
# Importing packages and libraries:
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
%matplotlib inline
import pylab as pl


# Importing the data:
hotel = pd.read_csv("hotel_bookings.csv")
np.set_printoptions(suppress=True, precision=5)
pd.set_option('display.precision', 5)
print(f"hotel dataframe has {hotel.shape[0]} rows and {hotel.shape[1]} columns.\nThe first 5 rows are:")
hotel.head()

hotel dataframe has 119390 rows and 32 columns.
The first 5 rows are:


# Five-number summary and beyond:
hotel.style.set_properties(**{"font-size":"11 px"})
hotel.describe(include = 'all').T


hotel.info()     # Showing the data types and number of non-null elements for each column

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 119390 entries, 0 to 119389
Data columns (total 32 columns):
 #   Column                          Non-Null Count   Dtype  
---  ------                          --------------   -----  
 0   hotel                           119390 non-null  object 
 1   is_canceled                     119390 non-null  int64  
 2   lead_time                       119390 non-null  int64  
 3   arrival_date_year               119390 non-null  int64  
 4   arrival_date_month              119390 non-null  object 
 5   arrival_date_week_number        119390 non-null  int64  
 6   arrival_date_day_of_month       119390 non-null  int64  
 7   stays_in_weekend_nights         119390 non-null  int64  
 8   stays_in_week_nights            119390 non-null  int64  
 9   adults                          119390 non-null  int64  
 10  children                        119386 non-null  float64
 11  babies                          119390 non-null  int64  
 12  meal                            119390 non-null  object 
 13  country                         118902 non-null  object 
 14  market_segment                  119390 non-null  object 
 15  distribution_channel            119390 non-null  object 
 16  is_repeated_guest               119390 non-null  int64  
 17  previous_cancellations          119390 non-null  int64  
 18  previous_bookings_not_canceled  119390 non-null  int64  
 19  reserved_room_type              119390 non-null  object 
 20  assigned_room_type              119390 non-null  object 
 21  booking_changes                 119390 non-null  int64  
 22  deposit_type                    119390 non-null  object 
 23  agent                           103050 non-null  float64
 24  company                         6797 non-null    float64
 25  days_in_waiting_list            119390 non-null  int64  
 26  customer_type                   119390 non-null  object 
 27  adr                             119390 non-null  float64
 28  required_car_parking_spaces     119390 non-null  int64  
 29  total_of_special_requests       119390 non-null  int64  
 30  reservation_status              119390 non-null  object 
 31  reservation_status_date         119390 non-null  object 
dtypes: float64(4), int64(16), object(12)
memory usage: 29.1+ MB


# Showing the columns containing Null values:
null_count = pd.DataFrame(hotel.isnull().sum()[hotel.isnull().sum()!=0], columns =["Null-count"])
print("Columns with Null values\n","-"*22,"\n", null_count)

Columns with Null values
 ---------------------- 
           Null-count
children           4
country          488
agent          16340
company       112593


hotel.dropna(subset = ["children"], inplace=True, axis = 0)   # Drop 4 rows containing Null in children column


# Showing the cancellation status counts among countries in a pivot table:

new_df = hotel[["country","is_canceled"]]
new_df = np.array(new_df)
dic_canc = {}
dic_notcan = {}

for row in new_df:                  # Iterate over rows of new_df
    if row[1]==1:                   # If canceled:
        if row[0] in dic_canc:           # If country exists in dic_canc:
            dic_canc[row[0]]+=1              # increment the value count
        else:
            dic_canc[row[0]]=1           # Otherwise add the country to the dic_canc
    if row[1]==0:                   # If not canceled:
        if row[0] in dic_notcan:         # If country exists in dic_notcan:
            dic_notcan[row[0]]+=1            # increment the value count
        else:
            dic_notcan[row[0]]=1         # Otherwise add the country to the dic_notcan
            
canceled_df = pd.DataFrame(dic_canc, index = ["is_canceled"]).T
notCanceled_df = pd.DataFrame(dic_notcan, index = ["not_canceled"]).T

pivot_df = pd.concat([canceled_df, notCanceled_df], axis=1)   # concat is outer join by default
pivot_df.fillna(0, inplace=True)      # fill the nulls generated as a result of concat (outer join)
pivot_df = pivot_df.astype(int)
pivot_df["total"] = pivot_df.sum(axis=1)
pivot_df["cancellation_percentage"]= 100 * pivot_df.is_canceled/pivot_df.total
print("\nThe mean cancellation percentage is", pivot_df.cancellation_percentage.mean())

# Any country occuring less than 500 times in the dataset is likely not much reliable to use for any prediction
high_cancel_country = pivot_df.query('total>500 & (cancellation_percentage>50)')
low_cancel_country = pivot_df.query('total>500 & (cancellation_percentage <20)')
low_cancel_country = low_cancel_country.sort_values(by=["total", "cancellation_percentage"], ascending = [False, True])
display(high_cancel_country)
display(low_cancel_country)

The mean cancellation percentage is 30.204589339202563


hotel["country"].fillna(value="unknown", inplace=True)      # Filling the 488 NA's with "unknown"
# Saving 'FRA','DEU','PRT', and 'unknown' in a np array to preserve them
pres = np.array(['FRA','DEU','PRT','unknown'])
# Replace any other country not in the pres list, by "listed_other":
hotel['country'] = hotel['country'].apply(lambda i: i if i in pres else 'listed_other')

# Now we only have a few unique values is the country column, shown below with counts:
display(pd.crosstab(hotel["country"], hotel["is_canceled"]))


print("Normalized value counts (percentages):\n","-"*35)
print(hotel.country.value_counts(normalize=True)*100)

Normalized value counts (percentages):
 -----------------------------------
listed_other    44.06714
PRT             40.69656
FRA              8.72380
DEU              6.10373
unknown          0.40876
Name: country, dtype: float64


# Showing the cancellation status counts among agents in a pivot table:

new_df = hotel[["agent","is_canceled"]].copy()
new_df.agent = (new_df.agent.astype('Int64')).astype(str)
new_df = np.array(new_df)

dic_canc = {}
dic_notcan = {}

for row in new_df:                  # Iterate over rows of new_df
    if row[1]==1:                   # If canceled:
        if row[0] in dic_canc:           # If agent exists in dic_canc:
            dic_canc[row[0]]+=1              # increment the value count
        else:
            dic_canc[row[0]]=1           # Otherwise add the agent to the dic_canc
    if row[1]==0:                   # If not canceled:
        if row[0] in dic_notcan:         # If agent exists in dic_notcan:
            dic_notcan[row[0]]+=1            # increment the value count
        else:
            dic_notcan[row[0]]=1         # Otherwise add the agent to the dic_notcan
            
canceled_df = pd.DataFrame(dic_canc, index = ["is_canceled"]).T
notCanceled_df = pd.DataFrame(dic_notcan, index = ["not_canceled"]).T

pivot_df = pd.concat([canceled_df, notCanceled_df], axis=1)   # concat is outer join by default
pivot_df.fillna(0, inplace=True)      # fill the nulls generated as a result of concat (outer join)
pivot_df = pivot_df.astype(int)
pivot_df["total"] = pivot_df.sum(axis=1)
pivot_df["cancellation_percentage"]= 100 * pivot_df.is_canceled/pivot_df.total
print("The mean cancellation percentage is", pivot_df.cancellation_percentage.mean())

# Any agent occuring less than 500 times in the dataset is likely not much reliable to use for prediction
high_cancel_agent = pivot_df.query('total>500 & (cancellation_percentage>50)')
high_cancel_agent = high_cancel_agent.sort_values(by=["total", "cancellation_percentage"], ascending = [False, False])
low_cancel_agent = pivot_df.query('total>500 & (cancellation_percentage <15)')
low_cancel_agent = low_cancel_agent.sort_values(by=["total", "cancellation_percentage"], ascending = [False, True])

display(high_cancel_agent)
display(low_cancel_agent)

The mean cancellation percentage is 25.187683521644328


hotel["agent"].fillna(value=-1.0, inplace=True)      # Filling the 16340 NA's with -1.0
hotel["agent"] = hotel["agent"].astype(int)
hotel["agent"] = hotel["agent"].astype(str)
hotel.loc[hotel["agent"] == "-1", "agent"] = "unknown"
# Saving unknown, "1", "unknown" in a np array to preserve them
pres = np.array(["unknown", "1"])
# Replace any other agent not in the pres list, by "listed_other":
hotel['agent'] = hotel['agent'].apply(lambda i: i if i in pres else 'listed_other')

# Now we only have a few unique values is the agent column, shown below with counts:
display(pd.crosstab(hotel["agent"], hotel["is_canceled"]))


print("Normalized value counts (percentages) for unique agents:\n","-"*53)
print(hotel.agent.value_counts(normalize=True)*100)

Normalized value counts (percentages) for unique agents:
 -----------------------------------------------------
listed_other    80.29166
unknown         13.68502
1                6.02332
Name: agent, dtype: float64


# Showing the cancellation status counts among companies in a pivot table:

new_df = hotel[["company","is_canceled"]].copy()
new_df.company = (new_df.company.astype('Int64')).astype(str)
new_df = np.array(new_df)

dic_canc = {}
dic_notcan = {}

for row in new_df:                  # Iterate over rows of new_df
    if row[1]==1:                   # If canceled:
        if row[0] in dic_canc:           # If company exists in dic_canc:
            dic_canc[row[0]]+=1              # increment the value count
        else:
            dic_canc[row[0]]=1           # Otherwise add the company to the dic_canc
    if row[1]==0:                   # If not canceled:
        if row[0] in dic_notcan:         # If company exists in dic_notcan:
            dic_notcan[row[0]]+=1            # increment the value count
        else:
            dic_notcan[row[0]]=1         # Otherwise add the company to the dic_notcan
            
canceled_df = pd.DataFrame(dic_canc, index = ["is_canceled"]).T
notCanceled_df = pd.DataFrame(dic_notcan, index = ["not_canceled"]).T

pivot_df = pd.concat([canceled_df, notCanceled_df], axis=1)   # concat is outer join by default
pivot_df.fillna(0, inplace=True)      # fill the nulls generated as a result of concat (outer join)
pivot_df = pivot_df.astype(int)
pivot_df["total"] = pivot_df.sum(axis=1)
pivot_df["cancellation_percentage"]= 100 * pivot_df.is_canceled/pivot_df.total
print("The mean cancellation percentage is", pivot_df.cancellation_percentage.mean())

# Any company occuring less than 300 times in the dataset is likely not much reliable to use for prediction
high_cancel_co = pivot_df.query('total>400 & (cancellation_percentage>50)')
high_cancel_co = high_cancel_co.sort_values(by=["total", "cancellation_percentage"], ascending = [False, False])
low_cancel_co = pivot_df.query('total>300 & (cancellation_percentage <5)')
low_cancel_co = low_cancel_co.sort_values(by=["total", "cancellation_percentage"], ascending = [False, True])

display(high_cancel_co)
display(low_cancel_co)

The mean cancellation percentage is 13.41212281412652


#company has 94% missing values
hotel.drop(columns = 'company', inplace=True)


#After handling missing values of all 4 columns (children, country, agent, company)
# Checking if any columns still containing Null values:
null_count = pd.DataFrame(hotel.isnull().sum()[hotel.isnull().sum()!=0], columns =["Null-count"])
print("Columns with Null values\n","-"*22,"\n", null_count)

Columns with Null values
 ---------------------- 
 Empty DataFrame
Columns: [Null-count]
Index: []


# No Nulls left in the dataframe.


# meal: Undefined (no meal), SC(no meal), BB(1 meal), HB(2 meals), FB(3 meals)
print("Meal unique values are: " , hotel.meal.unique())

# combining Undefined and SC because they mean the same thing based on the description (=no meal):
hotel.meal.replace("Undefined", "SC", inplace=True)
print("Meal unique values after modification: ", hotel.meal.unique())

Meal unique values are:  ['BB' 'FB' 'HB' 'SC' 'Undefined']
Meal unique values after modification:  ['BB' 'FB' 'HB' 'SC']


hotel.shape   # 4 rows containing null children was dropped earlier 119390 -4 = 119386

(119386, 31)


# Checking the categorical columns (object type):
hotel[hotel.columns[np.array(hotel.dtypes==object)]].head()


# Changing the reservation_status_date from object to date:
hotel.reservation_status_date = pd.to_datetime(hotel.reservation_status_date)


# checking the numerical columns:
hotel.select_dtypes(include=['int64', 'float64']).head()


# Children column type is float; changing the type to int:
hotel.children = hotel.children.astype(int)


# Ignoring rows where number of guests in total is zero:

baddata = (hotel.children == 0) & (hotel.adults == 0) & (hotel.babies == 0)
print(f"Rows with zero guests: {hotel[baddata].shape[0]}")
hotel = hotel[~baddata]
print(f"Rows left after ignoring rows with zero guests: {hotel.shape[0]}")

Rows with zero guests: 180
Rows left after ignoring rows with zero guests: 119206


# Checking the relation between status and other variables:
pd.set_option('display.max_rows', None)
hotel.groupby("reservation_status").describe().T     
# is_canceled column is similar to reservation_status, but with no-show and cancel cases combined and replaced by 1


# Information inferred from the above table:

# Reservations that get canceled: lead-time is high, previous cancellations are high, higher days in waitlist, etc.
# Reservations where guest stayed: low average of previous cancellations, higher number of special requests, etc.


# Reservation status count visualization
status_percentages = hotel["reservation_status"].value_counts(normalize=True)*100
print("Status percentages:","\n","-"*20)
print(status_percentages)

Status percentages: 
 --------------------
Check-Out    62.92552
Canceled     36.06278
No-Show       1.01169
Name: reservation_status, dtype: float64


status_percentages.plot(kind='bar', figsize=(4,3), ylabel="percentage");


# Relation between hotel type and cancellations:
pd.crosstab(hotel.hotel,hotel.is_canceled).plot.bar(figsize=(4,3), ylabel="count");
plt.show()


# Relation between deposit type and cancellations:
pd.crosstab(hotel.deposit_type,hotel.is_canceled).plot.bar(figsize=(4,3), ylabel="count");
plt.show()


# Relation between reservation status (counts of 1's for canceled and counts of 0's for not canceled) and arrival month
hotel.groupby(["arrival_date_month","is_canceled"])["is_canceled"].count().plot(kind='bar',figsize=(5,3));


# Checking if any of the independent numerical variables are highly correlated or are describing the same thing!
corrMatrix = hotel.corr()
corrMatrix.style.background_gradient(cmap='coolwarm')
# Some accidental correlations are observed (like stay in week vs weekend days) but overall, 
# numerical independent variables are not highly correlated.


corrMatrix["is_canceled"].sort_values(ascending=False)

is_canceled                       1.00000
lead_time                         0.29293
previous_cancellations            0.11015
adults                            0.05816
days_in_waiting_list              0.05431
adr                               0.04656
stays_in_week_nights              0.02555
arrival_date_year                 0.01669
arrival_date_week_number          0.00830
children                          0.00486
stays_in_weekend_nights          -0.00132
arrival_date_day_of_month        -0.00590
babies                           -0.03257
previous_bookings_not_canceled   -0.05736
is_repeated_guest                -0.08374
booking_changes                  -0.14482
required_car_parking_spaces      -0.19570
total_of_special_requests        -0.23493
Name: is_canceled, dtype: float64


# Variables lead_time and total_of_special_requests have the highest positive and negative 
# correlation with reservation status (is or is not canceled), respectively


# Relation between lead time and cancellations:
hotel.groupby(["is_canceled"])["lead_time"].mean()  # on average, canceled bookings have higher mean lead_time

is_canceled
0     80.08203
1    144.89911
Name: lead_time, dtype: float64


ct1 = pd.crosstab(hotel.total_of_special_requests, hotel.is_canceled)
ct1.plot(kind="bar", stacked=True, figsize=(5,3))   # The more the # of requests, the less the likelihood of cancellation

<AxesSubplot:xlabel='total_of_special_requests'>


# Checking the relation between arrival year and cancellations
hotel.groupby(['arrival_date_year', 'is_canceled']).size().to_frame(name='percentage').groupby(level=0).apply(lambda x:100 * x / float(x.sum()))


# Checking the relation between arrival month and cancellations
hotel.groupby(['arrival_date_month', 'is_canceled']).size().to_frame(name='percentage').groupby(level=0).apply(lambda x:100 * x / float(x.sum()))


# Cancellations doesn't seem to be very highly dependent on the arrival year or month.


# Listing the unique values in each categorical column
categorical_hotel = hotel[hotel.columns[np.array(hotel.dtypes==object)]]

for col in categorical_hotel:
  print(f"`{col}` column unique values:\n {hotel[col].unique()}\n")

`hotel` column unique values:
 ['Resort Hotel' 'City Hotel']

`arrival_date_month` column unique values:
 ['July' 'August' 'September' 'October' 'November' 'December' 'January'
 'February' 'March' 'April' 'May' 'June']

`meal` column unique values:
 ['BB' 'FB' 'HB' 'SC']

`country` column unique values:
 ['PRT' 'listed_other' 'FRA' 'unknown' 'DEU']

`market_segment` column unique values:
 ['Direct' 'Corporate' 'Online TA' 'Offline TA/TO' 'Complementary' 'Groups'
 'Aviation']

`distribution_channel` column unique values:
 ['Direct' 'Corporate' 'TA/TO' 'Undefined' 'GDS']

`reserved_room_type` column unique values:
 ['C' 'A' 'D' 'E' 'G' 'F' 'H' 'L' 'B']

`assigned_room_type` column unique values:
 ['C' 'A' 'D' 'E' 'G' 'F' 'I' 'B' 'H' 'L' 'K']

`deposit_type` column unique values:
 ['No Deposit' 'Refundable' 'Non Refund']

`agent` column unique values:
 ['unknown' 'listed_other' '1']

`customer_type` column unique values:
 ['Transient' 'Contract' 'Transient-Party' 'Group']

`reservation_status` column unique values:
 ['Check-Out' 'Canceled' 'No-Show']


hotel.head(3)


# Subtract the reservation_status_date which is the date for most recent status update, from
# the arrival date to better identify whether a canceled case would directly translate to 
# loss of profit or not

# First, it is required to concatenate the arrival columns (year, month, day) and generate an arrival_date column
import calendar
import datetime

arrival_Df = hotel[["arrival_date_year", "arrival_date_month", "arrival_date_day_of_month"]].copy()
arrival_Df.head()
months = list(calendar.month_name)


for month in months:
    arrival_Df.loc[arrival_Df["arrival_date_month"] == month, "arrival_date_month"] = months.index(month)
    
arrival_Df = arrival_Df.astype({'arrival_date_month': 'int64'})
arrival_Df.rename(columns = {'arrival_date_year':'year', 'arrival_date_month':'month', 'arrival_date_day_of_month':'day'}, inplace = True)
arri_date = pd.to_datetime(arrival_Df)
arrival_Df["arrival_date"] = arri_date

display(arrival_Df.head())
arrival_Df.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 119206 entries, 0 to 119389
Data columns (total 4 columns):
 #   Column        Non-Null Count   Dtype         
---  ------        --------------   -----         
 0   year          119206 non-null  int64         
 1   month         119206 non-null  int64         
 2   day           119206 non-null  int64         
 3   arrival_date  119206 non-null  datetime64[ns]
dtypes: datetime64[ns](1), int64(3)
memory usage: 4.5 MB


# Temporarily adding these new columns to the hotel dataframe:
hotel["arrival_date"] = arri_date
hotel['Difference'] = (hotel['arrival_date'] - hotel['reservation_status_date']).dt.days

# Now, generating a rule for what is profit and what can be a loss of profit, based on the date differences:
# If the reservation is canceled and the deposit is non-refundable or canceled way ahead of arrival (over 200 days),
# then the hotel gains profit by forfeiting the deposit
# If the guest has already checked out, then it's a profit case!
# If it is a no-show and the deposit is non-refundable, then it's a profit case!
profit = hotel.query('(reservation_status=="Canceled" & (deposit_type == "Non Refund" | Difference > 200)) | reservation_status=="Check-Out" | (reservation_status=="No-Show" & deposit_type == "Non Refund")').copy()

# Loss is a collection of instances not captured by the above query
loss = hotel.loc[hotel.index.difference(profit.index)].copy()

# The shapes row numbers add up to the total rows 119,206
print(profit.shape)
print(loss.shape)

display(profit.head())
display(loss.head())

(91487, 33)
(27719, 33)


# Creating a Revenue column for profit and loss dataframes and populating them:
profit["Revenue"] = 1
loss["Revenue"] = 0
hotel_2 = pd.concat([profit, loss], ignore_index=True)  
hotel_2 = hotel_2.sample(frac=1).reset_index(drop=True)  # Shuffle and reset index so profit and loss rows get shuffled
hotel_2.shape

(119206, 34)


hotel_2.head(3)


hotel_2.columns

Index(['hotel', 'is_canceled', 'lead_time', 'arrival_date_year',
       'arrival_date_month', 'arrival_date_week_number',
       'arrival_date_day_of_month', 'stays_in_weekend_nights',
       'stays_in_week_nights', 'adults', 'children', 'babies', 'meal',
       'country', 'market_segment', 'distribution_channel',
       'is_repeated_guest', 'previous_cancellations',
       'previous_bookings_not_canceled', 'reserved_room_type',
       'assigned_room_type', 'booking_changes', 'deposit_type', 'agent',
       'days_in_waiting_list', 'customer_type', 'adr',
       'required_car_parking_spaces', 'total_of_special_requests',
       'reservation_status', 'reservation_status_date', 'arrival_date',
       'Difference', 'Revenue'],
      dtype='object')


# drop these columns to avoid information leakage becase the target is created using these columns:
hotel_2.drop(columns = ['is_canceled', 'reservation_status_date', 'arrival_date', 'Difference', 'reservation_status'], inplace=True)
hotel_2.columns

Index(['hotel', 'lead_time', 'arrival_date_year', 'arrival_date_month',
       'arrival_date_week_number', 'arrival_date_day_of_month',
       'stays_in_weekend_nights', 'stays_in_week_nights', 'adults', 'children',
       'babies', 'meal', 'country', 'market_segment', 'distribution_channel',
       'is_repeated_guest', 'previous_cancellations',
       'previous_bookings_not_canceled', 'reserved_room_type',
       'assigned_room_type', 'booking_changes', 'deposit_type', 'agent',
       'days_in_waiting_list', 'customer_type', 'adr',
       'required_car_parking_spaces', 'total_of_special_requests', 'Revenue'],
      dtype='object')


hotel_2.head(3)


revenue_percentages = hotel_2["Revenue"].value_counts(normalize=True)*100
print(revenue_percentages)
revenue_percentages.plot(kind='bar', figsize=(4,3), ylabel="percentage");

1    76.74698
0    23.25302
Name: Revenue, dtype: float64


# Create dummy variables
hotel_ssf = pd.get_dummies(hotel_2)
print(f"There are now {hotel_ssf.shape[1]} columns in hotel_ssf dataframe.")

There are now 83 columns in hotel_ssf dataframe.


# Among the important independent variables that are correlated to Revenue are:


corrMatrix = hotel_ssf.corr()
corrMatrix["Revenue"].sort_values(ascending=False)[:5]

Revenue                         1.00000
deposit_type_Non Refund         0.20554
required_car_parking_spaces     0.14034
market_segment_Offline TA/TO    0.11960
market_segment_Groups           0.10615
Name: Revenue, dtype: float64


corrMatrix["Revenue"].sort_values(ascending=False)[-5:]

distribution_channel_TA/TO   -0.09863
adr                          -0.12158
agent_listed_other           -0.12832
deposit_type_No Deposit      -0.20466
market_segment_Online TA     -0.24340
Name: Revenue, dtype: float64


pd.crosstab(hotel_ssf["deposit_type_Non Refund"],hotel_ssf.Revenue).plot.bar(figsize=(4,3), ylabel="count");
plt.show()


pd.crosstab(hotel_ssf["deposit_type_No Deposit"],hotel_ssf.Revenue).plot.bar(figsize=(4,3), ylabel="count");
plt.show()


pd.crosstab(hotel_ssf["market_segment_Online TA"],hotel_ssf.Revenue).plot.bar(figsize=(4,3), ylabel="count");
plt.show()


# Data:    << hotel_ssf>>    is all numeric (includes dummy variables)
# Recall that hotel_2 still includes categorical variables that are not converted to dummies

# separate the Revenue column to be used as the target:
y = hotel_ssf["Revenue"]

X_ssf = hotel_ssf.drop(["Revenue"], axis=1)


# Creating a reduced version of the matrix X and label y to test run the classifiers easily

hotel_ssf_redu = hotel_ssf.sample(frac=0.12).reset_index(drop=True) #using only a fraction of rows
y = hotel_ssf_redu["Revenue"]
X_ssf = hotel_ssf_redu.drop(["Revenue"], axis=1)


revenue_percentages = hotel_ssf_redu["Revenue"].value_counts(normalize=True)*100
print(revenue_percentages)
revenue_percentages.plot(kind='bar', figsize=(4,3), ylabel="percentage");

1    76.28102
0    23.71898
Name: Revenue, dtype: float64


from sklearn.model_selection import train_test_split
#Using 80/20 split
train, test, target_train, target_test = train_test_split(X_ssf, y, test_size=0.2, random_state=33) 

from sklearn import preprocessing
#Normalizing the data
min_max_scaler = preprocessing.MinMaxScaler().fit(train)
#Min-Max normalization to scale all the variables between 0 & 1
train_norm = min_max_scaler.transform(train)
train_norm = pd.DataFrame(train_norm, columns=train.columns, index=train.index)

test_norm = min_max_scaler.transform(test)
test_norm = pd.DataFrame(test_norm, columns=test.columns, index=test.index)

target = y

train = train_norm
test= test_norm


train.shape

(11444, 82)


test.shape

(2861, 82)


from sklearn.model_selection import GridSearchCV
from sklearn import neighbors
#Intializing KNN classifier
clf = neighbors.KNeighborsClassifier()    
#Parameters for grid search (values of k=1, k=3, k=5, k=7, k=9)
parameters = {'n_neighbors': [1,3,5,7,9]}          
#Initializing the grid search with specified parameters & 5-fold cross-validation
gs = GridSearchCV(clf, parameters, verbose=1, cv=5)


tar_train = target_train.astype('int32')


tar_test = target_test.astype('int32')


#Performing the grid search on the training data
gs.fit(train, tar_train)

Fitting 5 folds for each of 5 candidates, totalling 25 fits

GridSearchCV(cv=5, estimator=KNeighborsClassifier(),
             param_grid={'n_neighbors': [1, 3, 5, 7, 9]}, verbose=1)


for (i, j) in gs.best_params_.items():
    print ("The optimal value of", i, "is:", j)
print()
print("The best cross-validation accuracy on the training data was: {}".format(gs.best_score_))

The optimal value of n_neighbors is: 7

The best cross-validation accuracy on the training data was: 0.7637194609671674


from sklearn.metrics import classification_report
#Initializing with k=7
clf = neighbors.KNeighborsClassifier(n_neighbors=7)  
#Fitting the training data 
clf.fit(train, tar_train) 
#Predict classes of test set
pred_test_target = clf.predict(test)   
#Print classification report
print(classification_report(tar_test, pred_test_target))

              precision    recall  f1-score   support

           0       0.55      0.42      0.48       702
           1       0.82      0.89      0.85      2159

    accuracy                           0.77      2861
   macro avg       0.69      0.65      0.66      2861
weighted avg       0.76      0.77      0.76      2861


from sklearn.model_selection import cross_val_score

clf = neighbors.KNeighborsClassifier(n_neighbors=3)
cv_scores = cross_val_score(clf4, train, tar_train, cv=10)

print(f"\nCross validation scores:\n{cv_scores}")
print("\n Overall Accuracy on X-Val: %0.2f (+/- %0.2f)" % (cv_scores.mean(), cv_scores.std() * 2))

clf.fit(train, tar_train)
print("\n Accuracy on Training: ",  clf4.score(train, tar_train))

clf.fit(test, tar_test)
print("\n Accuracy on Testing: ",  clf4.score(test, tar_test))

Cross validation scores:
[0.76157 0.78865 0.78253 0.79127 0.77797 0.78147 0.80332 0.77448 0.79458
 0.79808]

 Overall Accuracy on X-Val: 0.79 (+/- 0.02)

 Accuracy on Training:  0.932278923453338

 Accuracy on Testing:  0.787836420831877


# Splitting the data into train(80%) and test(20%) sets:
from sklearn.model_selection import train_test_split

train1, test1, target_train1, target_test1 = train_test_split(X_ssf, y, test_size=0.2, random_state=33)


from sklearn import tree
treeclf = tree.DecisionTreeClassifier(criterion='entropy', min_samples_split=3)


treeclf = treeclf.fit(train1, target_train1)


treepreds_test = treeclf.predict(test1)
print(treepreds_test)

[0 0 1 ... 1 1 1]


from sklearn.metrics import classification_report, accuracy_score, recall_score, precision_score, confusion_matrix, roc_curve, roc_auc_score,plot_confusion_matrix


print(classification_report(target_test1, treepreds_test))

              precision    recall  f1-score   support

           0       0.59      0.62      0.60       702
           1       0.87      0.86      0.87      2159

    accuracy                           0.80      2861
   macro avg       0.73      0.74      0.74      2861
weighted avg       0.80      0.80      0.80      2861


treecm = confusion_matrix(target_test1, treepreds_test)
print(treecm)

[[ 432  270]
 [ 298 1861]]


import pylab as plt
%matplotlib inline
plt.matshow(treecm)
plt.title('Confusion matrix')
plt.colorbar()
plt.ylabel('Actual')
plt.xlabel('Predicted')
plt.show()


pr = treeclf.predict_proba(test1)


import sklearn.metrics as metrics
prd = pr[:,1]
FalsePR, TruePR, threshold = metrics.roc_curve(target_test1, prd)
auc = metrics.auc(FalsePR, TruePR)


plt.title('ROC Curve')
plt.plot(FalsePR, TruePR, 'b', label = 'AUC = %0.2f' % auc)
plt.legend(loc = 'lower right')
plt.plot([0, 1], [0, 1],'r--')
plt.ylabel('TPR')
plt.xlabel('FPR')
plt.show()


treepreds_train = treeclf.predict(train1)
print(treepreds_train)

[1 1 0 ... 1 1 1]


print(classification_report(target_train1, treepreds_train))

              precision    recall  f1-score   support

           0       0.97      1.00      0.98      2691
           1       1.00      0.99      1.00      8753

    accuracy                           0.99     11444
   macro avg       0.99      0.99      0.99     11444
weighted avg       0.99      0.99      0.99     11444


treecm1 = confusion_matrix(target_train1, treepreds_train)
print(treecm1)

[[2685    6]
 [  76 8677]]


print (treeclf.score(train1, target_train1))

0.9928346731911919


print (treeclf.score(test1, target_test1))

0.8014680181754631


#Grid Search
from sklearn.model_selection import GridSearchCV

dt = tree.DecisionTreeClassifier()

parameters = {
    'criterion': ['entropy','gini'],
    'max_depth': np.linspace(1, 20, 10, dtype=int),
    'min_samples_leaf': np.linspace(1, 30, 15, dtype=int),
    'min_samples_split': np.linspace(2, 20, 10, dtype=int)
}

gs = GridSearchCV(dt, parameters, verbose=1, cv=5)


%time _ = gs.fit(train1, target_train1)

gs.best_params_, gs.best_score_

Fitting 5 folds for each of 3000 candidates, totalling 15000 fits
CPU times: user 8min 5s, sys: 1.19 s, total: 8min 6s
Wall time: 8min 7s

({'criterion': 'entropy',
  'max_depth': 11,
  'min_samples_leaf': 17,
  'min_samples_split': 4},
 0.8218290501547383)


from sklearn import metrics

def measure_performance(X, y, clf, show_accuracy=True, show_classification_report=True, show_confussion_matrix=True):
    y_pred = clf.predict(X)   
    if show_accuracy:
         print("Accuracy:{0:.3f}".format(metrics.accuracy_score(y, y_pred)),"\n")
    if show_classification_report:
        print("Classification report")
        print(metrics.classification_report(y, y_pred),"\n")
      
    if show_confussion_matrix:
        print("Confussion matrix")
        print(metrics.confusion_matrix(y, y_pred),"\n")


dtGS = tree.DecisionTreeClassifier(criterion='entropy', max_depth=11, min_samples_leaf=17, min_samples_split=4)

dtGS.fit(train1, target_train1)
measure_performance(test1, target_test1, dtGS, show_confussion_matrix=False, show_classification_report=True)

Accuracy:0.810 

Classification report
              precision    recall  f1-score   support

           0       0.63      0.54      0.58       702
           1       0.86      0.90      0.88      2159

    accuracy                           0.81      2861
   macro avg       0.74      0.72      0.73      2861
weighted avg       0.80      0.81      0.80      2861


from sklearn.model_selection import cross_val_score

treeclf = tree.DecisionTreeClassifier(criterion='entropy', max_depth=11, min_samples_leaf=17, min_samples_split=4)
cv_scores = cross_val_score(treeclf, train1, target_train1, cv=10)

print(f"\nCross validation scores:\n{cv_scores}")
print("\n Overall Accuracy on X-Val: %0.2f (+/- %0.2f)" % (cv_scores.mean(), cv_scores.std() * 2))

treeclf = treeclf.fit(train1, target_train1)
print("\n Accuracy on Training: ",  treeclf.score(train1, target_train1))

treeclf = treeclf.fit(test1, target_test1)
print("\n Accuracy on Testing: ",  treeclf.score(test1, target_test1))

Cross validation scores:
[0.81048 0.82707 0.81135 0.80961 0.81294 0.81469 0.81031 0.82168 0.8278
 0.83129]

 Overall Accuracy on X-Val: 0.82 (+/- 0.02)

 Accuracy on Training:  0.8405277874868927

 Accuracy on Testing:  0.8276826284515904


from sklearn.tree import export_graphviz
export_graphviz(treeclf,out_file='tree.dot', feature_names=train1.columns)


import graphviz

with open("tree.dot") as f:
    dot_graph = f.read()
graphviz.Source(dot_graph, format = "png")


dot_data = tree.export_graphviz(treeclf,out_file=None,feature_names=train1.columns,class_names=True)
graph = graphviz.Source(dot_data)
graph.render("image",view=True)

'image.pdf'


from sklearn.svm import SVC
# initialize svm classifier with balanced class weights
clf4 = SVC(class_weight='balanced', tol=1e-2, cache_size= 600)                  
# grid search parameters
parameters = {'gamma': [0.1, .25, .4, .5, .7], 'C': [20, 40, 50, 75, 100], 'kernel': ['rbf', 'linear']}
# initialize grid search with 3-fold cv
gs4 = GridSearchCV(clf4, parameters, verbose=1, cv=3)


#Perform grid search on sample
gs4.fit(train, target_train)

Fitting 3 folds for each of 50 candidates, totalling 150 fits

GridSearchCV(cv=3,
             estimator=SVC(cache_size=600, class_weight='balanced', tol=0.01),
             param_grid={'C': [20, 40, 50, 75, 100],
                         'gamma': [0.1, 0.25, 0.4, 0.5, 0.7],
                         'kernel': ['rbf', 'linear']},
             verbose=1)


for (i, j) in gs4.best_params_.items():
    print ("The optimal value of", i, "is:", j)
print()
print("The best cross-validation accuracy on the sample data was: {}".format(gs4.best_score_))

The optimal value of C is: 100
The optimal value of gamma is: 0.4
The optimal value of kernel is: rbf

The best cross-validation accuracy on the sample data was: 0.7862638234936336


#Initializing svm classifier with optimal parameters
clf4 = SVC(kernel='rbf', gamma =.4, C=100, class_weight='balanced') 
#Fit on training data
clf4.fit(train, target_train)   
#Predict classes of test set
pred_test_target = clf4.predict(test)
#Printing the classification report
print(classification_report(target_test, pred_test_target))  # print classification report

              precision    recall  f1-score   support

           0       0.55      0.68      0.61       702
           1       0.89      0.82      0.85      2159

    accuracy                           0.79      2861
   macro avg       0.72      0.75      0.73      2861
weighted avg       0.81      0.79      0.79      2861


from sklearn.model_selection import cross_val_score

clf4 = SVC(kernel='rbf', gamma =.7, C=100, class_weight='balanced') 
cv_scores = cross_val_score(clf4, train, target_train, cv=5)

print(f"\nCross validation scores:\n{cv_scores}")
print("\n Overall Accuracy on X-Val: %0.2f (+/- %0.2f)" % (cv_scores.mean(), cv_scores.std() * 2))

clf4.fit(train, target_train) 
print("\n Accuracy on Training: ",  clf4.score(train, target_train))

clf4.fit(test, target_test)
print("\n Accuracy on Testing: ",  clf4.score(test, target_test))

Cross validation scores:
[0.77632 0.78681 0.7903  0.782   0.79065]

 Overall Accuracy on X-Val: 0.79 (+/- 0.01)

 Accuracy on Training:  0.9529010835372247

 Accuracy on Testing:  0.9849702901083537


# Building a Naive Bayes classifier:
    
    # Notes:
    # 1-GaussianNB assumes distributions are normal 
    # 2-normalization is not needed for this classifier

from sklearn import naive_bayes
from sklearn.model_selection import cross_val_score  # for cross-validation
from sklearn.metrics import classification_report
    
nbclf = naive_bayes.GaussianNB()  
nbclf = nbclf.fit(train, target_train)
print("Score on Training: ", nbclf.score(train, target_train))
print("Score on Test: ", nbclf.score(test, target_test))

# perform 10-fold cross-validation on the 80% training data
cv_scores = cross_val_score(nbclf, train, target_train, cv=10)
print(f"\nCross validation scores:\n{cv_scores}")

# report the overall average accuracy.
print("\nOverall Accuracy: %0.2f (+/- %0.2f)" % (cv_scores.mean(),cv_scores.std()*2))

# Compare the cv accuracy to the model accuracy on the train data as a whole. 
print("\nAccuracy on Training: ",  nbclf.score(train, target_train))

# run your model on the set-aside 20% test data
nbpreds_test = nbclf.predict(test)
print("\nAccuracy on 20% test set: ",  nbclf.score(test, target_test))
print("\n\nclassification report:\n")
print(classification_report(target_test, nbpreds_test))

Score on Training:  0.4556099265990912
Score on Test:  0.4627752534078993

Cross validation scores:
[0.45066 0.45328 0.44192 0.44891 0.46241 0.44843 0.46766 0.46154 0.45717
 0.46416]

Overall Accuracy: 0.46 (+/- 0.02)

Accuracy on Training:  0.4556099265990912

Accuracy on 20% test set:  0.4627752534078993


classification report:

              precision    recall  f1-score   support

           0       0.31      0.97      0.47       702
           1       0.97      0.30      0.45      2159

    accuracy                           0.46      2861
   macro avg       0.64      0.64      0.46      2861
weighted avg       0.81      0.46      0.46      2861


# Hyperparameter tuning for Naive Bayes:
from sklearn.model_selection import GridSearchCV

parameters = {'var_smoothing': (1e-9,1e-7, 1e-5, 1e-4,1e-2)}

gs = GridSearchCV(nbclf, parameters, scoring = 'accuracy', cv=10)


_ = gs.fit(train, target_train)

gs.best_params_, gs.best_score_

({'var_smoothing': 0.01}, 0.5263936391119797)


# Creating the best model based on the above grid search results:

best_nbclf = naive_bayes.GaussianNB(var_smoothing= 0.01)  
best_nbclf = best_nbclf.fit(train, target_train)
print("Score on Training: ", best_nbclf.score(train, target_train))
print("Score on Test: ", best_nbclf.score(test, target_test))

# perform 10-fold cross-validation on the 80% training data
cv_scores = cross_val_score(best_nbclf, train, target_train, cv=10)
print(f"\nCross validation scores:\n{cv_scores}")

# report the overall average accuracy.
print("\nOverall Accuracy: %0.2f (+/- %0.2f)" % (cv_scores.mean(),cv_scores.std()*2))

# Compare the cv accuracy to the model accuracy on the train data as a whole. 
print("\nAccuracy on Training: ",  best_nbclf.score(train, target_train))

# run your model on the set-aside 20% test data
best_nbclf_test = best_nbclf.predict(test)
print("\nAccuracy on 20% test set: ",  best_nbclf.score(test, target_test))
print("\n\nclassification report:\n")
print(classification_report(target_test, best_nbclf_test))

Score on Training:  0.5297972736805313
Score on Test:  0.5456134218804614

Cross validation scores:
[0.50917 0.51703 0.52664 0.50393 0.54983 0.53584 0.53846 0.52448 0.52273
 0.53584]

Overall Accuracy: 0.53 (+/- 0.03)

Accuracy on Training:  0.5297972736805313

Accuracy on 20% test set:  0.5456134218804614


classification report:

              precision    recall  f1-score   support

           0       0.34      0.87      0.48       702
           1       0.91      0.44      0.59      2159

    accuracy                           0.55      2861
   macro avg       0.62      0.66      0.54      2861
weighted avg       0.77      0.55      0.57      2861


# Usin the grid search results (var_smoothing= 1e-07) slightly increased the accuracy on train and test


from sklearn.metrics import confusion_matrix
# generate the confusion matrix
nbcm = confusion_matrix(target_test, best_nbclf_test)
print(nbcm)
plt.matshow(nbcm)
plt.title('Confusion matrix')
plt.colorbar()
plt.ylabel('Actual')
plt.xlabel('Predicted')
plt.show();

[[ 611   91]
 [1209  950]]


nbmclf = naive_bayes.MultinomialNB()   
nbmclf = nbclf.fit(train, target_train)

# 10-fold cross-validation on the training data
cv_scores = cross_val_score(nbmclf, train, target_train, cv=10)
print(f"\nCross validation scores:\n{cv_scores}")

# Overall average accuracy.
print("\nOverall Accuracy: %0.2f (+/- %0.2f)" % (cv_scores.mean(),cv_scores.std()*2))

# Compare the cv accuracy to the model accuracy on the train data as a whole. 
print("\nAccuracy on Training: ",  nbmclf.score(train, target_train))

# run your model on the set-aside test data
nbmclf_test = nbmclf.predict(test)
print("\nAccuracy on test set: ",  nbmclf.score(test, target_test))
print("\n\nclassification report:\n")
print(classification_report(target_test, nbmclf_test))

Cross validation scores:
[0.45066 0.45328 0.44192 0.44891 0.46241 0.44843 0.46766 0.46154 0.45717
 0.46416]

Overall Accuracy: 0.46 (+/- 0.02)

Accuracy on Training:  0.4556099265990912

Accuracy on test set:  0.4627752534078993


classification report:

              precision    recall  f1-score   support

           0       0.31      0.97      0.47       702
           1       0.97      0.30      0.45      2159

    accuracy                           0.46      2861
   macro avg       0.64      0.64      0.46      2861
weighted avg       0.81      0.46      0.46      2861


from sklearn.discriminant_analysis import LinearDiscriminantAnalysis   # it is a classifier (supervised)

ldclf = LinearDiscriminantAnalysis()
ldclf = ldclf.fit(train, target_train)
ldpreds_test = ldclf.predict(test)

# 10-fold cross-validation on the training data
cv_scores = cross_val_score(ldclf, train, target_train, cv=10)
print(f"\nCross validation scores:\n{cv_scores}")

# Overall average accuracy.
print("\nOverall Accuracy: %0.2f (+/- %0.2f)" % (cv_scores.mean(),cv_scores.std()*2))

# Compare the cv accuracy to the model accuracy on the train data as a whole. 
print("\nAccuracy on Training: ",  ldclf.score(train, target_train))

# run your model on the set-aside test data
ldclf_test = ldclf.predict(test)
print("\nAccuracy on test set: ",  ldclf.score(test, target_test))
print("\n\nclassification report:\n")
print(classification_report(target_test, ldclf_test))

Cross validation scores:
[0.78865 0.7869  0.78865 0.79476 0.77797 0.76486 0.78234 0.78234 0.79196
 0.7771 ]

Overall Accuracy: 0.78 (+/- 0.02)

Accuracy on Training:  0.7866130723523244

Accuracy on test set:  0.7790982174065012


classification report:

              precision    recall  f1-score   support

           0       0.59      0.32      0.41       702
           1       0.81      0.93      0.86      2159

    accuracy                           0.78      2861
   macro avg       0.70      0.62      0.64      2861
weighted avg       0.75      0.78      0.75      2861


# hyperparameter tuning for LDA:
# The most important parameter to tune in LDA is the solver:

parameters = {'solver': ['svd', 'lsqr']}  
gs = GridSearchCV(ldclf, parameters, scoring='accuracy', cv=10)
gs_res = gs.fit(train, target_train)

gs_res.best_params_, gs_res.best_score_

({'solver': 'svd'}, 0.7835525391638928)


# Checkig if auto Shrinkage combined with the two other solvers provides a higher accuracy:
# Side note: svd only works with shrinkage=None

ldclf_2 = LinearDiscriminantAnalysis(shrinkage='auto')
parameters_2 = {'solver': ('lsqr','eigen')} 

gs = GridSearchCV(ldclf_2, parameters_2, scoring='accuracy', cv=10)
gs_res = gs.fit(train, target_train)

gs_res.best_params_, gs_res.best_score_

({'solver': 'lsqr'}, 0.778136317830641)


# Creating the best model based on the above grid search results: svd results in a higher accuracy
# We get the same results as the model without parameter tuning because svd is the default solver
best_ldclf = LinearDiscriminantAnalysis(solver='svd')
best_ldclf = best_ldclf.fit(train, target_train)
ldpreds_test = best_ldclf.predict(test)

# 10-fold cross-validation on the training data
cv_scores = cross_val_score(best_ldclf, train, target_train, cv=10)
print(f"\nCross validation scores:\n{cv_scores}")

# Overall average accuracy.
print("\nOverall Accuracy: %0.2f (+/- %0.2f)" % (cv_scores.mean(),cv_scores.std()*2))

# Compare the cv accuracy to the model accuracy on the train data as a whole. 
print("\nAccuracy on Training: ",  best_ldclf.score(train, target_train))

# run your model on the set-aside test data
best_ldclf_test = best_ldclf.predict(test)
print("\nAccuracy on test set: ",  best_ldclf.score(test, target_test))
print("\n\nclassification report:\n")
print(classification_report(target_test, best_ldclf_test))

Cross validation scores:
[0.78865 0.7869  0.78865 0.79476 0.77797 0.76486 0.78234 0.78234 0.79196
 0.7771 ]

Overall Accuracy: 0.78 (+/- 0.02)

Accuracy on Training:  0.7866130723523244

Accuracy on test set:  0.7790982174065012


classification report:

              precision    recall  f1-score   support

           0       0.59      0.32      0.41       702
           1       0.81      0.93      0.86      2159

    accuracy                           0.78      2861
   macro avg       0.70      0.62      0.64      2861
weighted avg       0.75      0.78      0.75      2861


# Above is the same accuracy as the model with no parameter tuning because svd is the default solver.


# generate the confusion matrix
ldcm = confusion_matrix(target_test, best_ldclf_test)
print(ldcm)

[[ 222  480]
 [ 152 2007]]


from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import GridSearchCV
from sklearn.metrics import classification_report, accuracy_score, recall_score, precision_score, confusion_matrix, roc_curve, roc_auc_score,plot_confusion_matrix
import warnings
warnings.filterwarnings('ignore')


def grid_search(X_tr, X_te, y_tr, y_te, model, params, scoring='recall'):
    gs = GridSearchCV(estimator = model, param_grid = params, scoring = scoring, n_jobs=-1, cv=3)
    gs.fit(X_tr, y_tr)
    y_pred = gs.predict(X_te)
    print(f"{model}")
    print(f"Best parameter      : {gs.best_params_}")
    print(f"Test Accuracy Score : {accuracy_score(y_te, y_pred)}")
    print(f"Train Accuracy Score: {accuracy_score(y_tr, gs.predict(X_tr))}")
    print(f"Recall score        : {recall_score(y_te, y_pred)}")
    print(f"Classification Report \n{'-'*30}\n {classification_report(y_te, y_pred)}")
    return gs.best_params_


params = {
    'n_estimators':[100, 200, 300, 400, 500, 600],
    'criterion' : ['gini', 'entropy'],
    'max_depth' : [5, 10, 15, 20],
}
model = RandomForestClassifier(random_state=33)
rf_best = grid_search(train, test, target_train, target_test, model, params, scoring='accuracy')

RandomForestClassifier(random_state=33)
Best parameter      : {'criterion': 'entropy', 'max_depth': 20, 'n_estimators': 600}
Test Accuracy Score : 0.8413142257951766
Train Accuracy Score: 0.9580566235581964
Recall score        : 0.9467345993515517
Classification Report 
------------------------------
               precision    recall  f1-score   support

           0       0.76      0.52      0.62       702
           1       0.86      0.95      0.90      2159

    accuracy                           0.84      2861
   macro avg       0.81      0.73      0.76      2861
weighted avg       0.83      0.84      0.83      2861


rf = RandomForestClassifier(n_estimators=rf_best['n_estimators'], criterion=rf_best['criterion'], max_depth=rf_best['max_depth'], 
                                  random_state=33)


from sklearn.model_selection import cross_val_score

cv_scores = cross_val_score(rf, train, target_train, cv=10)
print(f"\nCross validation scores:\n{cv_scores}")
print("\n Overall Accuracy on X-Val: %0.2f (+/- %0.2f)" % (cv_scores.mean(), cv_scores.std() * 2))

rf = rf.fit(train, target_train)
print("\n Accuracy on Training: ",  rf.score(train, target_train))

rf = rf.fit(test, target_test)
print("\n Accuracy on Testing: ",  rf.score(test, target_test))

Cross validation scores:
[0.84192 0.8559  0.83668 0.85066 0.84441 0.86276 0.83916 0.84878 0.8479
 0.85839]

 Overall Accuracy on X-Val: 0.85 (+/- 0.02)

 Accuracy on Training:  0.9580566235581964

 Accuracy on Testing:  0.9954561342188046


plot_confusion_matrix(rf, test, target_test)
plt.rcParams["figure.figsize"] = (12,12)
plt.show()


# Min-Max normalization:

from sklearn.preprocessing import MinMaxScaler
scaler = MinMaxScaler()
X_norm = scaler.fit_transform(X_ssf)


# plot_silhouettes function from class example:
# https://nbviewer.org/url/bmobasher.com/Class/CSC478/Clustering.ipynb

def plot_silhouettes(data, clusters, metric='euclidean'):
    
    from matplotlib import cm
    from sklearn.metrics import silhouette_samples

    cluster_labels = np.unique(clusters)
    n_clusters = cluster_labels.shape[0]
    silhouette_vals = metrics.silhouette_samples(data, clusters, metric='euclidean')
    c_ax_lower, c_ax_upper = 0, 0
    cticks = []
    for i, k in enumerate(cluster_labels):
        c_silhouette_vals = silhouette_vals[clusters == k]
        c_silhouette_vals.sort()
        c_ax_upper += len(c_silhouette_vals)
        color = cm.jet(float(i) / n_clusters)
        pl.barh(range(c_ax_lower, c_ax_upper), c_silhouette_vals, height=1.0, 
                      edgecolor='none', color=color)

        cticks.append((c_ax_lower + c_ax_upper) / 2)
        c_ax_lower += len(c_silhouette_vals)
    
    silhouette_avg = np.mean(silhouette_vals)
    pl.axvline(silhouette_avg, color="red", linestyle="--") 
    pl.yticks(cticks, cluster_labels)
    pl.ylabel('Cluster')
    pl.xlabel('Silhouette coefficient')
    pl.tight_layout()
    pl.show();
    
    return


from sklearn.cluster import KMeans 
from sklearn import metrics
from sklearn.metrics import completeness_score, homogeneity_score
from sklearn.metrics import silhouette_samples

def cluster_analysis(data, column_names, target, plot_silhouettes, k=7):
    "A function that does clustering analysis on data"
    
    DF = pd.DataFrame(data, columns=column_names)
    kms = KMeans(n_clusters=k, max_iter=500, verbose=0, n_init=5, random_state=33) 
    kms.fit(DF)
    clusters = kms.predict(data)   
    clusters = clusters.astype(int)
    
    cluster_num, size = np.unique(clusters, return_counts=True)
    print("\n"+"-"*10+"Cluster Analysis"+"-"*10+"\n")
    print(f"Number of clusters: {k}")
    
    centroids = pd.DataFrame(kms.cluster_centers_, columns=column_names)
    silhouettes = metrics.silhouette_samples(data, clusters)
    
    for i in range(len(cluster_num)):
        print("Size of Cluster", cluster_num[i], "= ", size[i])
    
    print("\nCluster centroids:")
#     pd.set_option('max_columns', None)
    display(centroids)
    print("\nSilhouette plot:")
    plot_silhouettes(data, clusters)
    
    print("\n\nMean Silhouette Value: ", silhouettes.mean())
    complete = completeness_score(target,clusters)
    print("Completeness score: ", complete)
    homogen = homogeneity_score(target,clusters)
    print("Homogeneity score: ",homogen,"\n")
    
    return silhouettes, complete, homogen, clusters, centroids


silhouettes_mean, completeness , homogeneity  =[],[],[]
k_range= list(range(2,20))  # trying k=2 all the way to k=20
for K in k_range:   
    sil,comp,hmg,_,_=cluster_analysis(X_norm,X_ssf.columns,y,plot_silhouettes,k=K)
    silhouettes_mean.append(sil.mean())
    completeness.append(comp)
    homogeneity.append(hmg)

----------Cluster Analysis----------

Number of clusters: 2
Size of Cluster 0 =  6104
Size of Cluster 1 =  8201

Cluster centroids:

Silhouette plot:


Mean Silhouette Value:  0.120709553994198
Completeness score:  0.020741610562666327
Homogeneity score:  0.02583576722378519 


----------Cluster Analysis----------

Number of clusters: 3
Size of Cluster 0 =  2561
Size of Cluster 1 =  7629
Size of Cluster 2 =  4115

Cluster centroids:

Silhouette plot:


Mean Silhouette Value:  0.1334278627385949
Completeness score:  0.016093857919452847
Homogeneity score:  0.029426563177585862 


----------Cluster Analysis----------

Number of clusters: 4
Size of Cluster 0 =  3812
Size of Cluster 1 =  4263
Size of Cluster 2 =  2459
Size of Cluster 3 =  3771

Cluster centroids:

Silhouette plot:


Mean Silhouette Value:  0.11352242091907526
Completeness score:  0.011107883199360064
Homogeneity score:  0.02772500571421824 


----------Cluster Analysis----------

Number of clusters: 5
Size of Cluster 0 =  2485
Size of Cluster 1 =  4111
Size of Cluster 2 =  1637
Size of Cluster 3 =  3681
Size of Cluster 4 =  2391

Cluster centroids:

Silhouette plot:


Mean Silhouette Value:  0.1304795384489958
Completeness score:  0.02564862841455512
Homogeneity score:  0.07298119283366054 


----------Cluster Analysis----------

Number of clusters: 6
Size of Cluster 0 =  2023
Size of Cluster 1 =  2340
Size of Cluster 2 =  1762
Size of Cluster 3 =  2577
Size of Cluster 4 =  3256
Size of Cluster 5 =  2347

Cluster centroids:

Silhouette plot:


fig, ax = plt.subplots()
fig.set_size_inches(5, 5)
ax.plot(k_range, silhouettes_mean, label = "silhouette mean")
ax.plot(k_range, completeness, label="completeness")
ax.plot(k_range, homogeneity, label = "homogeneity")
plt.xlabel("K")
plt.legend(loc='lower right')
plt.show();


# Best K: 

# Based on the above scores plot, k=5, k=7, and k=2 have relatively higher scores (Sil.mean, completeness and homogeneity)
# compared to others; However, we'll look at centroids for k=2 to see if we can detect interesting patterns:


sil,comp,hmg,clust,cent=cluster_analysis(X_norm,X_ssf.columns,y,plot_silhouettes,k=2)

----------Cluster Analysis----------

Number of clusters: 2
Size of Cluster 0 =  6104
Size of Cluster 1 =  8201

Cluster centroids:

Silhouette plot:


Mean Silhouette Value:  0.120709553994198
Completeness score:  0.020741610562666327
Homogeneity score:  0.02583576722378519


# Analysis of Cluster centroids 0 and 1:

cent.T   # centroid 0: shows characteristics of profit; Centroid 1: loss


2.14746e-03

0.00214746

	count	unique	top	freq	first	last	mean	std	min	25%	50%	75%	max
hotel	119206	2	City Hotel	79159	NaT	NaT	NaN	NaN	NaN	NaN	NaN	NaN	NaN
is_canceled	119206.0	NaN	NaN	NaN	NaT	NaT	0.37074	0.48301	0.0	0.0	0.0	1.0	1.0
lead_time	119206.0	NaN	NaN	NaN	NaT	NaT	104.11262	106.87564	0.0	18.0	69.0	161.0	737.0
arrival_date_year	119206.0	NaN	NaN	NaN	NaT	NaT	2016.15651	0.70747	2015.0	2016.0	2016.0	2017.0	2017.0
arrival_date_month	119206	12	August	13857	NaT	NaT	NaN	NaN	NaN	NaN	NaN	NaN	NaN
arrival_date_week_number	119206.0	NaN	NaN	NaN	NaT	NaT	27.1632	13.6013	1.0	16.0	28.0	38.0	53.0
arrival_date_day_of_month	119206.0	NaN	NaN	NaN	NaT	NaT	15.79903	8.78102	1.0	8.0	16.0	23.0	31.0
stays_in_weekend_nights	119206.0	NaN	NaN	NaN	NaT	NaT	0.92706	0.99512	0.0	0.0	1.0	2.0	19.0
stays_in_week_nights	119206.0	NaN	NaN	NaN	NaT	NaT	2.4992	1.89711	0.0	1.0	2.0	3.0	50.0
adults	119206.0	NaN	NaN	NaN	NaT	NaT	1.85919	0.57519	0.0	2.0	2.0	2.0	55.0
children	119206.0	NaN	NaN	NaN	NaT	NaT	0.10405	0.39884	0.0	0.0	0.0	0.0	10.0
babies	119206.0	NaN	NaN	NaN	NaT	NaT	0.00796	0.09751	0.0	0.0	0.0	0.0	10.0
meal	119206	4	BB	92232	NaT	NaT	NaN	NaN	NaN	NaN	NaN	NaN	NaN
country	119206	5	listed_other	52563	NaT	NaT	NaN	NaN	NaN	NaN	NaN	NaN	NaN
market_segment	119206	7	Online TA	56407	NaT	NaT	NaN	NaN	NaN	NaN	NaN	NaN	NaN
distribution_channel	119206	5	TA/TO	97750	NaT	NaT	NaN	NaN	NaN	NaN	NaN	NaN	NaN
is_repeated_guest	119206.0	NaN	NaN	NaN	NaT	NaT	0.0315	0.17467	0.0	0.0	0.0	0.0	1.0
previous_cancellations	119206.0	NaN	NaN	NaN	NaT	NaT	0.08719	0.84493	0.0	0.0	0.0	0.0	26.0
previous_bookings_not_canceled	119206.0	NaN	NaN	NaN	NaT	NaT	0.1371	1.49816	0.0	0.0	0.0	0.0	72.0
reserved_room_type	119206	9	A	85873	NaT	NaT	NaN	NaN	NaN	NaN	NaN	NaN	NaN
assigned_room_type	119206	11	A	74020	NaT	NaT	NaN	NaN	NaN	NaN	NaN	NaN	NaN
booking_changes	119206.0	NaN	NaN	NaN	NaT	NaT	0.21881	0.63851	0.0	0.0	0.0	0.0	18.0
deposit_type	119206	3	No Deposit	104457	NaT	NaT	NaN	NaN	NaN	NaN	NaN	NaN	NaN
agent	119206	3	listed_other	95741	NaT	NaT	NaN	NaN	NaN	NaN	NaN	NaN	NaN
days_in_waiting_list	119206.0	NaN	NaN	NaN	NaT	NaT	2.32129	17.59829	0.0	0.0	0.0	0.0	391.0
customer_type	119206	4	Transient	89476	NaT	NaT	NaN	NaN	NaN	NaN	NaN	NaN	NaN
adr	119206.0	NaN	NaN	NaN	NaT	NaT	101.97152	50.43287	-6.38	69.5	94.95	126.0	5400.0
required_car_parking_spaces	119206.0	NaN	NaN	NaN	NaT	NaT	0.06256	0.24536	0.0	0.0	0.0	0.0	8.0
total_of_special_requests	119206.0	NaN	NaN	NaN	NaT	NaT	0.57148	0.79288	0.0	0.0	0.0	1.0	5.0
reservation_status	119206	3	Check-Out	75011	NaT	NaT	NaN	NaN	NaN	NaN	NaN	NaN	NaN
reservation_status_date	119206	926	2015-10-21 00:00:00	1460	2014-10-17	2017-09-14	NaN	NaN	NaN	NaN	NaN	NaN	NaN
arrival_date	119206	793	2015-12-05 00:00:00	448	2015-07-01	2017-08-31	NaN	NaN	NaN	NaN	NaN	NaN	NaN
Difference	119206.0	NaN	NaN	NaN	NaT	NaT	29.72146	70.10175	-69.0	-3.0	-1.0	26.0	526.0

	is_canceled	not_canceled	total	cancellation_percentage
FRA	1934	8481	10415	18.56937
DEU	1218	6069	7287	16.71470
NLD	387	1717	2104	18.39354
CN	254	1025	1279	19.85927
AUT	230	1033	1263	18.21061

	reservation_status	Canceled	Check-Out	No-Show
is_canceled	count	42989.00000	75011.00000	1206.00000
	mean	1.00000	0.00000	1.00000
	std	0.00000	0.00000	0.00000
	min	1.00000	0.00000	1.00000
	25%	1.00000	0.00000	1.00000
	50%	1.00000	0.00000	1.00000
	75%	1.00000	0.00000	1.00000
	max	1.00000	0.00000	1.00000
lead_time	count	42989.00000	75011.00000	1206.00000
	mean	147.35858	80.08203	57.22886
	std	118.81389	91.13780	66.63515
	min	0.00000	0.00000	0.00000
	25%	51.00000	9.00000	5.00000
	50%	116.00000	45.00000	30.00000
	75%	219.00000	124.00000	99.75000
	max	629.00000	737.00000	385.00000
arrival_date_year	count	42989.00000	75011.00000	1206.00000
	mean	2016.17307	2016.14745	2016.13018
	std	0.71610	0.70310	0.65600
	min	2015.00000	2015.00000	2015.00000
	25%	2016.00000	2016.00000	2016.00000
	50%	2016.00000	2016.00000	2016.00000
	75%	2017.00000	2017.00000	2017.00000
	max	2017.00000	2017.00000	2017.00000
arrival_date_week_number	count	42989.00000	75011.00000	1206.00000
	mean	27.40303	27.07656	24.00332
	std	13.01656	13.89714	14.88796
	min	1.00000	1.00000	1.00000
	25%	17.00000	16.00000	10.00000
	50%	28.00000	28.00000	22.00000
	75%	38.00000	38.00000	36.00000
	max	53.00000	53.00000	53.00000
arrival_date_day_of_month	count	42989.00000	75011.00000	1206.00000
	mean	15.69606	15.83881	16.99502
	std	8.77620	8.77649	9.12704
	min	1.00000	1.00000	1.00000
	25%	8.00000	8.00000	9.00000
	50%	16.00000	16.00000	18.00000
	75%	23.00000	23.00000	25.00000
	max	31.00000	31.00000	31.00000
stays_in_weekend_nights	count	42989.00000	75011.00000	1206.00000
	mean	0.92056	0.92806	1.09619
	std	1.00014	0.98782	1.22723
	min	0.00000	0.00000	0.00000
	25%	0.00000	0.00000	0.00000
	50%	1.00000	1.00000	1.00000
	75%	2.00000	2.00000	2.00000
	max	16.00000	19.00000	8.00000
stays_in_week_nights	count	42989.00000	75011.00000	1206.00000
	mean	2.56091	2.46200	2.61360
	std	1.84856	1.90742	2.72288
	min	0.00000	0.00000	0.00000
	25%	1.00000	1.00000	1.00000
	50%	2.00000	2.00000	2.00000
	75%	3.00000	3.00000	3.00000
	max	40.00000	50.00000	21.00000
adults	count	42989.00000	75011.00000	1206.00000
	mean	1.90947	1.83352	1.66418
	std	0.67866	0.50415	0.55337
	min	0.00000	0.00000	0.00000
	25%	2.00000	2.00000	1.00000
	50%	2.00000	2.00000	2.00000
	75%	2.00000	2.00000	2.00000
	max	55.00000	4.00000	3.00000
children	count	42989.00000	75011.00000	1206.00000
	mean	0.10624	0.10256	0.11857
	std	0.40879	0.39121	0.49733
	min	0.00000	0.00000	0.00000
	25%	0.00000	0.00000	0.00000
	50%	0.00000	0.00000	0.00000
	75%	0.00000	0.00000	0.00000
	max	3.00000	3.00000	10.00000
babies	count	42989.00000	75011.00000	1206.00000
	mean	0.00375	0.01040	0.00663
	std	0.06184	0.11312	0.08121
	min	0.00000	0.00000	0.00000
	25%	0.00000	0.00000	0.00000
	50%	0.00000	0.00000	0.00000
	75%	0.00000	0.00000	0.00000
	max	2.00000	10.00000	1.00000
is_repeated_guest	count	42989.00000	75011.00000	1206.00000
	mean	0.01175	0.04273	0.03731
	std	0.10775	0.20224	0.18961
	min	0.00000	0.00000	0.00000
	25%	0.00000	0.00000	0.00000
	50%	0.00000	0.00000	0.00000
	75%	0.00000	0.00000	0.00000
	max	1.00000	1.00000	1.00000
previous_cancellations	count	42989.00000	75011.00000	1206.00000
	mean	0.21412	0.01576	0.00580
	std	1.35078	0.27249	0.10370
	min	0.00000	0.00000	0.00000
	25%	0.00000	0.00000	0.00000
	50%	0.00000	0.00000	0.00000
	75%	0.00000	0.00000	0.00000
	max	26.00000	13.00000	3.00000
previous_bookings_not_canceled	count	42989.00000	75011.00000	1206.00000
	mean	0.02261	0.20306	0.11526
	std	0.64554	1.81202	1.42909
	min	0.00000	0.00000	0.00000
	25%	0.00000	0.00000	0.00000
	50%	0.00000	0.00000	0.00000
	75%	0.00000	0.00000	0.00000
	max	58.00000	72.00000	44.00000
booking_changes	count	42989.00000	75011.00000	1206.00000
	mean	0.09456	0.28978	0.23300
	std	0.44322	0.71724	0.66040
	min	0.00000	0.00000	0.00000
	25%	0.00000	0.00000	0.00000
	50%	0.00000	0.00000	0.00000
	75%	0.00000	0.00000	0.00000
	max	16.00000	18.00000	6.00000
days_in_waiting_list	count	42989.00000	75011.00000	1206.00000
	mean	3.66647	1.58769	0.00000
	std	21.78664	14.78122	0.00000
	min	0.00000	0.00000	0.00000
	25%	0.00000	0.00000	0.00000
	50%	0.00000	0.00000	0.00000
	75%	0.00000	0.00000	0.00000
	max	391.00000	379.00000	0.00000
adr	count	42989.00000	75011.00000	1206.00000
	mean	105.27101	100.16921	96.45837
	std	52.70032	49.07064	44.96386
	min	0.00000	-6.38000	0.00000
	25%	73.00000	67.76000	69.50000
	50%	96.30000	92.70000	86.39500
	75%	127.98000	125.00000	118.80000
	max	5400.00000	510.00000	328.67000
required_car_parking_spaces	count	42989.00000	75011.00000	1206.00000
	mean	0.00000	0.09941	0.00000
	std	0.00000	0.30333	0.00000
	min	0.00000	0.00000	0.00000
	25%	0.00000	0.00000	0.00000
	50%	0.00000	0.00000	0.00000
	75%	0.00000	0.00000	0.00000
	max	0.00000	8.00000	0.00000
total_of_special_requests	count	42989.00000	75011.00000	1206.00000
	mean	0.32311	0.71446	0.53234
	std	0.64467	0.83403	0.76665
	min	0.00000	0.00000	0.00000
	25%	0.00000	0.00000	0.00000
	50%	0.00000	1.00000	0.00000
	75%	0.00000	1.00000	1.00000
	max	5.00000	5.00000	4.00000

	is_canceled	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	previous_cancellations	previous_bookings_not_canceled	booking_changes	days_in_waiting_list	adr	required_car_parking_spaces	total_of_special_requests	Difference
is_canceled	1.000000	0.292930	0.016694	0.008299	-0.005902	-0.001316	0.025549	0.058155	0.004862	-0.032566	-0.083740	0.110147	-0.057363	-0.144821	0.054309	0.046558	-0.195696	-0.234925	0.615310
lead_time	0.292930	1.000000	0.040285	0.127060	0.002272	0.085981	0.166892	0.117601	-0.037886	-0.021006	-0.123217	0.086023	-0.073603	0.002219	0.170007	-0.065068	-0.116634	-0.095924	0.560562
arrival_date_year	0.016694	0.040285	1.000000	-0.540378	-0.000179	0.021685	0.031198	0.030305	0.054698	-0.013197	0.010272	-0.119916	0.029231	0.031398	-0.056357	0.198367	-0.013826	0.108663	0.048352
arrival_date_week_number	0.008299	0.127060	-0.540378	1.000000	0.066587	0.018630	0.016048	0.026559	0.005559	0.010418	-0.031123	0.035494	-0.021008	0.006316	0.022679	0.076302	0.001983	0.026192	0.045317
arrival_date_day_of_month	-0.005902	0.002272	-0.000179	0.066587	1.000000	-0.016241	-0.028381	-0.001728	0.014541	-0.000238	-0.006478	-0.027031	-0.000309	0.011254	0.022528	0.030234	0.008561	0.003058	-0.016418
stays_in_weekend_nights	-0.001316	0.085981	0.021685	0.018630	-0.016241	1.000000	0.494173	0.094777	0.046134	0.018606	-0.086011	-0.012769	-0.042860	0.050190	-0.054401	0.050652	-0.018521	0.073139	-0.049712
stays_in_week_nights	0.025549	0.166892	0.031198	0.016048	-0.028381	0.494173	1.000000	0.096222	0.044651	0.020373	-0.095305	-0.013977	-0.048874	0.080018	-0.002026	0.066828	-0.024934	0.068745	0.016530
adults	0.058155	0.117601	0.030305	0.026559	-0.001728	0.094777	0.096222	1.000000	0.029416	0.017892	-0.140972	-0.007068	-0.108856	-0.041465	-0.008362	0.224302	0.014444	0.123324	0.067498
children	0.004862	-0.037886	0.054698	0.005559	0.014541	0.046134	0.044651	0.029416	1.000000	0.023999	-0.032477	-0.024755	-0.021079	0.050997	-0.033294	0.325058	0.056245	0.081756	-0.018232
babies	-0.032566	-0.021006	-0.013197	0.010418	-0.000238	0.018606	0.020373	0.017892	0.023999	1.000000	-0.008813	-0.007509	-0.006552	0.085605	-0.010627	0.029041	0.037389	0.097943	-0.026516
is_repeated_guest	-0.083740	-0.123217	0.010272	-0.031123	-0.006478	-0.086011	-0.095305	-0.140972	-0.032477	-0.008813	1.000000	0.082740	0.420642	0.013042	-0.022058	-0.130821	0.077926	0.012968	-0.036032
previous_cancellations	0.110147	0.086023	-0.119916	0.035494	-0.027031	-0.012769	-0.013977	-0.007068	-0.024755	-0.007509	0.082740	1.000000	0.152570	-0.027262	0.005941	-0.065982	-0.018541	-0.048485	0.113764
previous_bookings_not_canceled	-0.057363	-0.073603	0.029231	-0.021008	-0.000309	-0.042860	-0.048874	-0.108856	-0.021079	-0.006552	0.420642	0.152570	1.000000	0.011962	-0.009417	-0.072342	0.047506	0.037778	-0.038162
booking_changes	-0.144821	0.002219	0.031398	0.006316	0.011254	0.050190	0.080018	-0.041465	0.050997	0.085605	0.013042	-0.027262	0.011962	1.000000	-0.011918	0.026586	0.067487	0.055014	-0.113227
days_in_waiting_list	0.054309	0.170007	-0.056357	0.022679	0.022528	-0.054401	-0.002026	-0.008362	-0.033294	-0.010627	-0.022058	0.005941	-0.009417	-0.011918	1.000000	-0.040867	-0.030603	-0.082752	0.039739
adr	0.046558	-0.065068	0.198367	0.076302	0.030234	0.050652	0.066828	0.224302	0.325058	0.029041	-0.130821	-0.065982	-0.072342	0.026586	-0.040867	1.000000	0.056500	0.172361	-0.024432
required_car_parking_spaces	-0.195696	-0.116634	-0.013826	0.001983	0.008561	-0.018521	-0.024934	0.014444	0.056245	0.037389	0.077926	-0.018541	0.047506	0.067487	-0.030603	0.056500	1.000000	0.082726	-0.119616
total_of_special_requests	-0.234925	-0.095924	0.108663	0.026192	0.003058	0.073139	0.068745	0.123324	0.081756	0.097943	0.012968	-0.048485	0.037778	0.055014	-0.082752	0.172361	0.082726	1.000000	-0.202507
Difference	0.615310	0.560562	0.048352	0.045317	-0.016418	-0.049712	0.016530	0.067498	-0.018232	-0.026516	-0.036032	0.113764	-0.038162	-0.113227	0.039739	-0.024432	-0.119616	-0.202507	1.000000

		percentage
arrival_date_year	is_canceled
2015	0	62.95133
2015	1	37.04867
2016	0	64.10646
2016	1	35.89354
2017	0	61.26539
2017	1	38.73461

DePaul University Jarvis College of Computing and Digital Media¶

DSC 478 Final Project¶

Outline¶

1 - Data Collection:¶

1.1 - Dataset: Hotel booking demand ¶

1.2 - Importing the Data:¶

2 - Data Preprocessing:¶

2.1 - Summary Statistics:¶

2.2 - Data Cleaning:¶

2.3 - Data Visualization and Exploration:¶

2.4 - Feature Engineering:¶

I can now safely drop the columns:¶

2.5 Data Visualization for Target Variable¶

Preparing the data and labels for classification¶

3 - Supervised Knowledge Discovery¶

3.1 KNN¶

First split, then train the minmax scalar on train set only; then use it to transform test:¶

3.2 Decision Trees¶

3.3 SVM¶

3.4 Naive Bayes¶

3.5 linear discriminant analysis¶

3.6 Random Forest¶

4 - Unsupervised Knowledge Discovery¶

4.1 Clustering¶

4.2 Qualitative Analysis of Clusters¶

Detecting patterns among cluster centroids:¶

5 - Comparison and Conclusion¶

KNN : 0.80(+/- 0.01)¶

Decision Trees : 0.82 (+/- 0.02)¶

Random Forest : : 0.85 (+/- 0.02)¶

SVM: 0.79 (+/- 0.01)¶

LDA: 0.78 (+/- 0.02)¶

Naïve Bayes: 0.53 (+/- 0.03)¶

	hotel	lead_time	arrival_date_year	arrival_date_month	arrival_date_week_number	arrival_date_day_of_month	stays_in_week_nights	adults	...	deposit_type	agent	company	customer_type	adr	total_of_special_requests	reservation_status	reservation_status_date
0	Resort Hotel	342	2015	July	27	1	0	2	...	No Deposit	NaN	NaN	Transient	0.0	0	Check-Out	2015-07-01
1	Resort Hotel	737	2015	July	27	1	0	2	...	No Deposit	NaN	NaN	Transient	0.0	0	Check-Out	2015-07-01
2	Resort Hotel	7	2015	July	27	1	1	1	...	No Deposit	NaN	NaN	Transient	75.0	0	Check-Out	2015-07-02
3	Resort Hotel	13	2015	July	27	1	1	1	...	No Deposit	304.0	NaN	Transient	75.0	0	Check-Out	2015-07-02
4	Resort Hotel	14	2015	July	27	1	2	2	...	No Deposit	240.0	NaN	Transient	98.0	1	Check-Out	2015-07-03

	is_canceled	not_canceled	total	cancellation_percentage
1	5280	1911	7191	73.42511
3	771	565	1336	57.70958
37	717	513	1230	58.29268
19	780	281	1061	73.51555
21	506	369	875	57.82857
229	484	302	786	61.57761
29	546	137	683	79.94143
12	304	274	578	52.59516
20	359	181	540	66.48148

	is_canceled	not_canceled	total	cancellation_percentage
7	474	3065	3539	13.39361
241	236	1485	1721	13.71296
28	110	1556	1666	6.60264
40	84	955	1039	8.08470
243	28	486	514	5.44747

	hotel	arrival_date_month	meal	country	market_segment	distribution_channel	reserved_room_type	assigned_room_type	deposit_type	agent	customer_type	reservation_status	reservation_status_date
0	Resort Hotel	July	BB	PRT	Direct	Direct	C	C	No Deposit	unknown	Transient	Check-Out	2015-07-01
1	Resort Hotel	July	BB	PRT	Direct	Direct	C	C	No Deposit	unknown	Transient	Check-Out	2015-07-01
2	Resort Hotel	July	BB	listed_other	Direct	Direct	A	C	No Deposit	unknown	Transient	Check-Out	2015-07-02
3	Resort Hotel	July	BB	listed_other	Corporate	Corporate	A	A	No Deposit	listed_other	Transient	Check-Out	2015-07-02
4	Resort Hotel	July	BB	listed_other	Online TA	TA/TO	A	A	No Deposit	listed_other	Transient	Check-Out	2015-07-03

		percentage
arrival_date_month	is_canceled
April	0	59.21647
April	1	40.78353
August	0	62.23569
August	1	37.76431
December	0	64.96523
December	1	35.03477
February	0	66.55489
February	1	33.44511
January	0	69.49840
January	1	30.50160
July	0	62.53559
July	1	37.46441
June	0	58.51405
June	1	41.48595
March	0	67.77232
March	1	32.22768
May	0	60.29711
May	1	39.70289
November	0	68.69000
November	1	31.31000
October	0	61.90903
October	1	38.09097
September	0	60.80952
September	1	39.19048

	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	...	deposit_type_No Deposit	deposit_type_Non Refund	deposit_type_Refundable	agent_1	agent_listed_other	agent_unknown	customer_type_Contract	customer_type_Group	customer_type_Transient	customer_type_Transient-Party
0	0.17278	0.48648	0.52169	0.49641	0.04530	0.05394	0.06369	0.01092	0.00377	0.06193	...	0.71494	0.28244	0.00262	0.14007	0.55013	0.30980	0.03539	0.00508	0.56062	0.39892
1	0.11522	0.64181	0.48953	0.49122	0.06727	0.06863	0.07260	0.05361	0.00451	0.01097	...	0.99902	0.00085	0.00012	0.00012	0.98549	0.01439	0.03487	0.00573	0.90416	0.05524

	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	...	deposit_type_No Deposit	deposit_type_Non Refund	deposit_type_Refundable	agent_1	agent_listed_other	agent_unknown	customer_type_Contract	customer_type_Group	customer_type_Transient	customer_type_Transient-Party
0	0.07422	0.57536	0.49854	0.50138	0.04617	0.05489	0.06245	0.03918	0.00957	0.12612	...	0.96017	0.03592	0.00390	-6.93889e-18	0.34752	0.65248	0.00195	0.00976	0.79032	0.19797
1	0.11597	0.63618	0.49156	0.48884	0.06681	0.06832	0.07258	0.05158	0.00413	0.00983	...	0.99974	0.00013	0.00013	2.62261e-04	0.99607	0.00367	0.03383	0.00551	0.90454	0.05612
2	0.22464	0.46333	0.52783	0.49701	0.04867	0.05599	0.06575	0.00308	0.00097	0.01700	...	0.60078	0.39777	0.00146	2.07382e-01	0.71734	0.07528	0.05804	0.00267	0.46503	0.47426

	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	...	deposit_type_No Deposit	deposit_type_Non Refund	deposit_type_Refundable	agent_1	agent_listed_other	agent_unknown	customer_type_Contract	customer_type_Group	customer_type_Transient	customer_type_Transient-Party
0	0.23049	0.46392	0.52652	0.49815	0.04579	0.05357	0.06565	0.00271	0.00092	0.01679	...	0.57203	4.26397e-01	1.57439e-03	2.21202e-01	0.69877	0.08003	0.02781	0.00105	0.47442	0.49672
1	0.11818	0.64024	0.49343	0.48679	0.06319	0.06500	0.06894	0.01814	0.00328	0.00891	...	0.99977	-7.21645e-16	2.34522e-04	5.27356e-16	0.99578	0.00422	0.05535	0.00610	0.89329	0.04526
2	0.07233	0.57035	0.49914	0.50194	0.04532	0.05388	0.06204	0.03985	0.00956	0.12973	...	0.96218	3.37536e-02	4.06669e-03	-5.55112e-17	0.32574	0.67426	0.00203	0.00976	0.78406	0.20415
3	0.11651	0.61854	0.49352	0.49065	0.07234	0.07380	0.07622	0.08503	0.00504	0.01246	...	0.99390	6.09918e-03	2.16840e-18	3.44736e-03	0.98913	0.00743	0.04110	0.00636	0.87298	0.07955

	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	...	deposit_type_No Deposit	deposit_type_Non Refund	deposit_type_Refundable	agent_1	agent_listed_other	agent_unknown	customer_type_Contract	customer_type_Group	customer_type_Transient	customer_type_Transient-Party
0	0.19024	0.42696	0.54180	0.50168	0.05609	0.06003	0.06473	0.00550	1.20724e-03	0.02374	...	0.97787	2.01207e-02	2.01207e-03	1.52918e-01	0.78994	0.05714	0.05875	4.42656e-03	0.14286	0.79396
1	0.11382	0.64461	0.49250	0.48742	0.06060	0.06256	0.06894	0.01824	3.40550e-03	0.00851	...	0.99976	-7.91034e-16	2.43250e-04	2.43250e-04	0.99514	0.00462	0.03503	5.10825e-03	0.92654	0.03333
2	0.28447	0.53696	0.49554	0.48962	0.03955	0.05264	0.06690	0.00020	1.38778e-17	0.00611	...	0.00489	9.95113e-01	-2.81893e-18	2.88943e-01	0.57605	0.13500	0.03787	1.04083e-17	0.92242	0.03971
3	0.11518	0.61994	0.49338	0.49083	0.07238	0.07366	0.07640	0.08666	5.02581e-03	0.01222	...	0.99674	3.25998e-03	1.30104e-18	5.43331e-04	0.99321	0.00625	0.03939	6.24830e-03	0.88346	0.07090
4	0.07079	0.56922	0.50216	0.50185	0.04538	0.05371	0.06238	0.04113	1.02468e-02	0.13342	...	0.97867	1.67294e-02	4.60059e-03	-1.11022e-16	0.32915	0.67085	0.00209	9.61941e-03	0.79925	0.18904

	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	...	deposit_type_No Deposit	deposit_type_Non Refund	deposit_type_Refundable	agent_1	agent_listed_other	agent_unknown	customer_type_Contract	customer_type_Group	customer_type_Transient	customer_type_Transient-Party
0	0.20102	0.41024	0.54616	0.50396	0.04680	0.05261	0.06424	0.00445	7.41840e-04	0.01731	...	0.97329	2.42334e-02	2.47280e-03	1.87933e-01	0.74184	0.07023	0.01879	1.48368e-03	0.01632	0.96340
1	0.12078	0.56728	0.50499	0.49244	0.08260	0.08334	0.07086	0.06578	8.11619e-03	0.02179	...	1.00000	-5.13478e-16	-3.03577e-18	-1.45717e-16	0.99231	0.00769	0.05382	5.55318e-03	0.86758	0.07305
2	0.27101	0.53462	0.49629	0.49274	0.04054	0.05241	0.06691	0.00114	1.56125e-17	0.00908	...	0.07605	9.23950e-01	-2.60209e-18	2.68445e-01	0.60670	0.12486	0.03519	6.93889e-18	0.92679	0.03802
3	0.11784	0.64435	0.49196	0.49164	0.07007	0.07201	0.07793	0.06881	2.91036e-03	0.00893	...	0.99573	4.26853e-03	-3.03577e-18	7.76096e-04	0.99302	0.00621	0.04734	8.14901e-03	0.87776	0.06674
4	0.11080	0.65617	0.48892	0.48393	0.05595	0.05695	0.06860	0.02058	2.30344e-03	0.00768	...	0.99969	-6.93889e-16	3.07125e-04	3.07125e-04	0.99539	0.00430	0.04607	5.22113e-03	0.92045	0.02826
5	0.07175	0.56945	0.50206	0.50107	0.04516	0.05424	0.06233	0.04133	1.04389e-02	0.13549	...	0.97699	1.83213e-02	4.68683e-03	-1.31839e-16	0.31870	0.68130	0.00170	1.02258e-02	0.80145	0.18662

	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	...	deposit_type_No Deposit	deposit_type_Non Refund	deposit_type_Refundable	agent_1	agent_listed_other	agent_unknown	customer_type_Contract	customer_type_Group	customer_type_Transient	customer_type_Transient-Party
0	0.13706	0.60013	0.50337	0.49484	0.09166	0.09270	0.07434	0.09713	6.73941e-03	0.02054	...	0.99101	8.98588e-03	-2.60209e-18	6.41849e-04	0.98588	0.01348	0.05777	4.49294e-03	0.81900	0.11874
1	0.19650	0.40371	0.54800	0.50769	0.04649	0.05176	0.06403	0.00449	7.22195e-04	0.01926	...	0.97352	2.40732e-02	2.40732e-03	1.82956e-01	0.75060	0.06644	0.01637	9.62927e-04	0.06259	0.92008
2	0.11114	0.65967	0.48762	0.48372	0.05586	0.05689	0.06861	0.02029	2.28311e-03	0.00791	...	0.99970	-7.07767e-16	3.04414e-04	3.04414e-04	0.99543	0.00426	0.04505	5.17504e-03	0.92146	0.02831
3	0.11852	0.54554	0.51143	0.49087	0.08110	0.08172	0.06897	0.01155	8.01796e-03	0.02502	...	0.98974	1.02630e-02	-2.60209e-18	-2.35922e-16	0.99423	0.00577	0.07056	8.98012e-03	0.85119	0.06928
4	0.28518	0.54141	0.49336	0.49011	0.03905	0.05241	0.06684	0.00021	1.38778e-17	0.00618	...	0.00494	9.95056e-01	-2.81893e-18	2.92336e-01	0.57108	0.13659	0.03832	1.21431e-17	0.92151	0.04017
5	0.10839	0.66134	0.48532	0.49125	0.06047	0.06294	0.07942	0.08677	2.83505e-03	0.00361	...	0.99897	1.03093e-03	-2.81893e-18	5.15464e-04	0.99536	0.00412	0.02784	8.24742e-03	0.91186	0.05206
6	0.06893	0.56570	0.50160	0.49950	0.04338	0.05247	0.06183	0.04071	9.92063e-03	0.13845	...	0.97795	1.71958e-02	4.85009e-03	-1.73472e-16	0.29541	0.70459	0.00176	9.70018e-03	0.80115	0.18739

	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	...	deposit_type_No Deposit	deposit_type_Non Refund	deposit_type_Refundable	agent_1	agent_listed_other	agent_unknown	customer_type_Contract	customer_type_Group	customer_type_Transient	customer_type_Transient-Party
0	0.12062	0.56727	0.50575	0.49253	0.08266	0.08335	0.07084	0.06584	8.14053e-03	0.02228	...	1.00000	-4.99600e-16	-3.46945e-18	-1.38778e-16	0.99186	0.00814	0.05398	5.56984e-03	0.86718	0.07326
1	0.12275	0.65437	0.48994	0.49715	0.07233	0.07388	0.07844	0.06827	2.40175e-03	0.00961	...	0.99520	4.80349e-03	-3.46945e-18	4.36681e-04	0.99345	0.00611	0.04498	9.17031e-03	0.87424	0.07162
2	0.20587	0.40841	0.54739	0.50465	0.04656	0.05263	0.06448	0.00410	5.13084e-04	0.01642	...	0.97229	2.56542e-02	2.05233e-03	1.94459e-01	0.73730	0.06824	0.01591	5.13084e-04	0.01437	0.96921
3	0.11045	0.59439	0.50281	0.48681	0.05907	0.05914	0.06873	0.03209	3.13939e-03	0.01005	...	1.00000	-5.27356e-16	-3.03577e-18	1.25576e-03	0.99372	0.00502	0.06279	4.60444e-03	0.90038	0.03223
4	0.10460	0.72684	0.47029	0.47417	0.04862	0.05196	0.06925	0.00879	1.99681e-03	0.00958	...	0.99920	4.16334e-17	7.98722e-04	-1.94289e-16	0.99521	0.00479	0.02077	7.18850e-03	0.94089	0.03115
5	0.07047	0.55221	0.50358	0.48698	0.03228	0.04546	0.04889	0.00303	1.13507e-03	0.24972	...	0.93303	5.33485e-02	1.36209e-02	-1.31839e-16	0.06697	0.93303	0.00114	6.81044e-03	0.62316	0.36890
6	0.28733	0.52921	0.50167	0.49293	0.03963	0.05244	0.06724	0.00062	1.30104e-17	0.00615	...	0.00492	9.95080e-01	-2.60209e-18	2.90898e-01	0.58057	0.12854	0.03813	1.12757e-17	0.92128	0.04059
7	0.07045	0.59186	0.49267	0.50495	0.05303	0.05945	0.06958	0.06124	1.48359e-02	0.06061	...	0.99684	3.15657e-03	-2.81893e-18	-2.35922e-16	0.49874	0.50126	0.00189	1.07323e-02	0.88826	0.09912

	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	...	deposit_type_No Deposit	deposit_type_Non Refund	deposit_type_Refundable	agent_1	agent_listed_other	agent_unknown	customer_type_Contract	customer_type_Group	customer_type_Transient	customer_type_Transient-Party
0	0.07075	0.59265	0.49258	0.50560	0.05288	0.05963	0.06965	0.06113	1.50160e-02	0.06070	...	9.96805e-01	3.19489e-03	-2.81893e-18	-2.49800e-16	0.50607	0.49393	0.00192	1.08626e-02	0.89457	0.09265
1	0.11100	0.59467	0.50269	0.48993	0.05844	0.05886	0.06895	0.04386	3.00842e-03	0.00963	...	1.00000e+00	-5.41234e-16	-3.25261e-18	-3.46945e-17	0.99479	0.00521	0.06097	5.21460e-03	0.89130	0.04252
2	0.12375	0.65510	0.49046	0.49429	0.07347	0.07459	0.07858	0.05729	2.29779e-03	0.01011	...	9.94945e-01	5.05515e-03	-3.25261e-18	4.59559e-04	0.99357	0.00597	0.04274	8.27206e-03	0.87500	0.07399
3	0.20630	0.37500	0.53750	0.51362	0.05093	0.04968	0.06511	0.00270	-7.80626e-18	0.02778	...	9.17824e-01	7.52315e-02	6.94444e-03	5.25463e-01	0.33102	0.14352	0.02431	1.15741e-03	0.01389	0.96065
4	0.28389	0.55344	0.49646	0.49257	0.03996	0.05296	0.06704	0.00064	1.21431e-17	0.00580	...	4.44089e-16	1.00000e+00	-2.60209e-18	2.58210e-01	0.60786	0.13393	0.03992	1.47451e-17	0.95943	0.00064
5	0.07099	0.54989	0.50495	0.48843	0.03305	0.04579	0.04936	0.00443	1.10865e-03	0.24390	...	9.36807e-01	5.21064e-02	1.10865e-02	-1.38778e-16	0.06541	0.93459	0.00111	5.54324e-03	0.61530	0.37805
6	0.21178	0.42408	0.55504	0.49381	0.05094	0.06125	0.06481	0.00446	1.57356e-03	0.01259	...	9.60661e-01	3.93391e-02	-2.38524e-18	-2.01228e-16	0.98899	0.01101	0.05193	3.93391e-03	0.06845	0.87569
7	0.11640	0.56906	0.50370	0.49329	0.07999	0.08093	0.07102	0.06846	8.29596e-03	0.02063	...	1.00000e+00	-4.85723e-16	-3.03577e-18	-1.94289e-16	0.99327	0.00673	0.03543	4.48430e-03	0.89193	0.06816
8	0.10468	0.72742	0.47012	0.47437	0.04866	0.05192	0.06925	0.00879	1.99840e-03	0.00959	...	9.99201e-01	4.16334e-17	7.99361e-04	-2.08167e-16	0.99520	0.00480	0.01998	7.19424e-03	0.94165	0.03118

	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	...	deposit_type_No Deposit	deposit_type_Non Refund	deposit_type_Refundable	agent_1	agent_listed_other	agent_unknown	customer_type_Contract	customer_type_Group	customer_type_Transient	customer_type_Transient-Party
0	0.21542	0.59013	0.46357	0.48798	0.04159	0.05478	0.06620	2.39163e-03	4.48430e-04	0.01525	...	0.13004	8.69955e-01	-2.16840e-18	-1.87350e-16	9.79372e-01	0.02063	3.46945e-17	1.12757e-17	0.99821	0.00179
1	0.12191	0.65241	0.49141	0.49861	0.07229	0.07374	0.07848	6.90058e-02	2.41228e-03	0.00965	...	0.99605	3.94737e-03	-3.25261e-18	4.38596e-04	9.95614e-01	0.00395	4.56140e-02	9.21053e-03	0.87807	0.06711
2	0.11671	0.56918	0.50368	0.49305	0.08154	0.08257	0.07089	6.79970e-02	8.42572e-03	0.02129	...	1.00000	-4.71845e-16	-3.25261e-18	-1.94289e-16	9.92905e-01	0.00710	3.72506e-02	3.99113e-03	0.89401	0.06475
3	0.05789	0.53915	0.51385	0.47445	0.03193	0.04323	0.04788	3.42319e-03	6.41849e-04	0.26316	...	0.94095	5.90501e-02	-1.30104e-18	-9.02056e-17	8.08729e-02	0.91913	1.28370e-03	6.41849e-03	0.68164	0.31065
4	0.10466	0.72764	0.47019	0.47463	0.04862	0.05194	0.06925	8.78594e-03	1.99681e-03	0.00958	...	0.99920	4.16334e-17	7.98722e-04	-1.94289e-16	9.95208e-01	0.00479	1.99681e-02	7.18850e-03	0.94169	0.03115
5	0.20058	0.46516	0.50670	0.51288	0.05339	0.05561	0.06482	3.01659e-03	4.52489e-04	0.02443	...	0.98281	2.71493e-03	1.44796e-02	3.45701e-01	4.62443e-01	0.19186	1.99095e-02	1.80995e-03	0.00271	0.97557
6	0.06929	0.59187	0.49218	0.50646	0.05211	0.05886	0.06931	6.04167e-02	1.46875e-02	0.06812	...	0.99687	3.12500e-03	-2.81893e-18	-2.35922e-16	4.90000e-01	0.51000	1.87500e-03	1.06250e-02	0.89250	0.09500
7	0.21042	0.38521	0.57871	0.49657	0.04569	0.05632	0.06429	4.80307e-03	1.44092e-03	0.01153	...	0.95485	4.51489e-02	-1.95156e-18	-1.73472e-16	9.87512e-01	0.01249	5.76369e-02	5.76369e-03	0.05476	0.88184
8	0.36742	0.42413	0.56586	0.50577	0.03907	0.04818	0.06907	4.16334e-17	-6.93889e-18	0.00152	...	0.01214	9.87860e-01	-6.50521e-19	7.17754e-01	1.11022e-16	0.28225	9.40819e-02	-6.07153e-18	0.80728	0.09863
9	0.11195	0.60005	0.50214	0.48433	0.05935	0.05950	0.06877	3.34986e-02	2.92925e-03	0.00676	...	1.00000	-4.85723e-16	-3.25261e-18	-1.87350e-16	9.94592e-01	0.00541	6.35421e-02	4.05588e-03	0.89049	0.04191

	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	...	deposit_type_No Deposit	deposit_type_Non Refund	deposit_type_Refundable	agent_1	agent_listed_other	agent_unknown	customer_type_Contract	customer_type_Group	customer_type_Transient	customer_type_Transient-Party
0	0.21245	0.37931	0.58243	0.50087	0.04023	0.05248	0.06301	0.00453	1.56740e-03	0.00731	...	0.94984	5.01567e-02	-1.95156e-18	-1.59595e-16	0.98746	0.01254	0.01567	5.20417e-18	0.04493	9.39394e-01
1	0.11467	0.63913	0.49495	0.48905	0.05962	0.05978	0.06873	0.09749	2.70819e-03	0.00677	...	1.00000	-1.38778e-16	-2.60209e-18	-2.15106e-16	0.99255	0.00745	0.03656	3.38524e-03	0.90995	5.01016e-02
2	0.11893	0.54757	0.51138	0.49167	0.08078	0.08142	0.06909	0.01165	7.76699e-03	0.02654	...	0.98964	1.03560e-02	-2.81893e-18	-2.42861e-16	0.99353	0.00647	0.06926	9.70874e-03	0.85307	6.79612e-02
3	0.10362	0.72705	0.46868	0.47502	0.04818	0.05155	0.06914	0.00662	1.65426e-03	0.00993	...	0.99917	6.93889e-17	8.27130e-04	-1.87350e-16	0.99504	0.00496	0.02068	7.44417e-03	0.93962	3.22581e-02
4	0.05817	0.54021	0.51319	0.47289	0.03186	0.04316	0.04770	0.00346	6.48508e-04	0.26459	...	0.94034	5.96628e-02	-1.30104e-18	-9.71445e-17	0.07912	0.92088	0.00130	6.48508e-03	0.67834	3.13878e-01
5	0.06933	0.59114	0.49310	0.50643	0.05220	0.05900	0.06933	0.06055	1.46692e-02	0.06804	...	0.99688	3.12110e-03	-2.60209e-18	-2.35922e-16	0.49064	0.50936	0.00187	1.06117e-02	0.89014	9.73783e-02
6	0.28454	0.55834	0.49396	0.49259	0.03940	0.05271	0.06698	0.00022	1.21431e-17	0.00587	...	0.00065	9.99348e-01	-2.60209e-18	2.61408e-01	0.60235	0.13625	0.04042	1.47451e-17	0.95958	-3.88578e-16
7	0.11048	0.55534	0.50801	0.49636	0.05761	0.05848	0.06942	0.05112	3.55731e-03	0.01265	...	0.99921	7.90514e-04	-2.38524e-18	-2.08167e-16	0.99130	0.00870	0.08538	7.11462e-03	0.86482	4.26877e-02
8	0.12442	0.64381	0.49376	0.48929	0.07557	0.07616	0.07907	0.01281	2.40128e-03	0.01067	...	0.99413	5.86980e-03	-3.03577e-18	5.33618e-04	0.99360	0.00587	0.04589	9.07150e-03	0.86820	7.68410e-02
9	0.20900	0.42496	0.51500	0.51368	0.05112	0.05373	0.06509	0.00178	4.44050e-04	0.02398	...	0.92540	6.03908e-02	1.42096e-02	4.03197e-01	0.41918	0.17762	0.01865	1.12757e-17	0.00888	9.72469e-01
10	0.12400	0.62804	0.48964	0.49552	0.08241	0.08426	0.07631	0.14956	7.40741e-03	0.01376	...	0.99683	3.17460e-03	-1.95156e-18	-1.45717e-16	0.98624	0.01376	0.02116	1.05820e-03	0.89312	8.46561e-02

	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	...	deposit_type_No Deposit	deposit_type_Non Refund	deposit_type_Refundable	agent_1	agent_listed_other	agent_unknown	customer_type_Contract	customer_type_Group	customer_type_Transient	customer_type_Transient-Party
0	0.13225	0.60241	0.51257	0.49885	0.08763	0.08795	0.07254	6.87751e-02	7.15361e-03	0.00828	...	0.99096	9.03614e-03	-2.60209e-18	-2.15106e-16	9.92470e-01	7.53012e-03	8.35843e-02	3.76506e-03	8.13253e-01	0.09940
1	0.11243	0.61558	0.49753	0.49130	0.05856	0.05979	0.06976	8.13073e-02	2.98924e-03	0.00638	...	0.99960	3.98565e-04	-3.46945e-18	1.19570e-03	9.92826e-01	5.97848e-03	6.37704e-02	5.57991e-03	8.91989e-01	0.03866
2	0.20135	0.51669	0.52676	0.47112	0.04192	0.05665	0.06621	5.19095e-03	5.56174e-04	0.02336	...	0.33148	6.68521e-01	-1.51788e-18	-1.31839e-16	9.74416e-01	2.55840e-02	4.44939e-03	3.46945e-18	9.42158e-01	0.05339
3	0.11256	0.67167	0.48480	0.48785	0.06362	0.06420	0.08201	1.21753e-02	2.02922e-03	0.00325	...	0.99919	8.11688e-04	-2.38524e-18	-2.01228e-16	9.97565e-01	2.43506e-03	1.29870e-02	9.74026e-03	9.22890e-01	0.05438
4	0.05861	0.54461	0.51055	0.46866	0.03212	0.04254	0.04764	3.10697e-03	6.65779e-04	0.26498	...	0.94407	5.59254e-02	-1.30104e-18	-9.71445e-17	7.58988e-02	9.24101e-01	1.33156e-03	6.65779e-03	6.77763e-01	0.31425
5	0.05780	0.52685	0.49029	0.51178	0.03974	0.05150	0.06608	3.68779e-02	9.66702e-03	0.11171	...	0.96778	2.14823e-02	1.07411e-02	-1.38778e-16	1.11022e-16	1.00000e+00	-2.08167e-17	4.29646e-03	7.99141e-01	0.19656
6	0.10577	0.72604	0.47226	0.47516	0.04897	0.05226	0.06934	8.78594e-03	1.99681e-03	0.00879	...	0.99920	4.16334e-17	7.98722e-04	-1.94289e-16	9.96006e-01	3.99361e-03	2.07668e-02	7.18850e-03	9.40895e-01	0.03115
7	0.20366	0.41730	0.54424	0.50811	0.04801	0.05218	0.06384	3.98984e-03	8.16104e-04	0.01578	...	0.99565	1.08814e-03	3.26442e-03	2.06202e-01	7.24701e-01	6.90968e-02	1.68662e-02	5.44070e-04	-8.88178e-16	0.98259
8	0.31114	0.52927	0.48321	0.51008	0.04001	0.05043	0.06867	9.02056e-17	-2.60209e-18	0.00096	...	0.00768	9.92322e-01	-1.73472e-18	4.53935e-01	3.79079e-01	1.66987e-01	5.95010e-02	1.04083e-17	8.78119e-01	0.06238
9	0.10663	0.51665	0.49770	0.48240	0.07591	0.07713	0.06914	6.26836e-02	9.30460e-03	0.04310	...	0.99412	5.87659e-03	-2.16840e-18	-1.66533e-16	9.89226e-01	1.07738e-02	1.37120e-02	9.79432e-03	9.31440e-01	0.04505
10	0.15129	0.59494	0.50903	0.49297	0.09687	0.09761	0.07314	1.31271e-02	2.81294e-03	0.02391	...	0.98312	1.68776e-02	-1.08420e-18	1.40647e-03	9.76090e-01	2.25035e-02	1.04079e-01	7.03235e-03	7.36990e-01	0.15190
11	0.09726	0.65613	0.49830	0.51475	0.06329	0.06745	0.07159	7.96460e-02	1.89633e-02	0.01391	...	1.00000	1.11022e-16	-1.30104e-18	-1.11022e-16	1.00000e+00	8.32667e-17	3.79267e-03	1.64349e-02	9.02655e-01	0.07712

	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	...	deposit_type_No Deposit	deposit_type_Non Refund	deposit_type_Refundable	agent_1	agent_listed_other	agent_unknown	customer_type_Contract	customer_type_Group	customer_type_Transient	customer_type_Transient-Party
0	0.27752	0.58259	0.48371	0.49511	0.04014	0.05312	0.06689	0.00023	9.54098e-18	0.00617	...	6.66134e-16	1.00000e+00	-2.81893e-18	2.24812e-01	0.63331	0.14188	1.17961e-16	1.64799e-17	1.00000	-3.88578e-16
1	0.06909	0.59213	0.49417	0.50766	0.05178	0.05881	0.06943	0.06163	1.50352e-02	0.06270	...	9.97441e-01	2.55918e-03	-2.81893e-18	-2.35922e-16	0.49904	0.50096	1.91939e-03	1.08765e-02	0.90467	8.25336e-02
2	0.11231	0.67073	0.48577	0.48835	0.06387	0.06429	0.08205	0.01220	2.03252e-03	0.00325	...	9.99187e-01	8.13008e-04	-2.16840e-18	-2.08167e-16	0.99675	0.00325	1.30081e-02	9.75610e-03	0.92764	4.95935e-02
3	0.10575	0.73043	0.47049	0.47571	0.04863	0.05184	0.06929	0.00834	2.01776e-03	0.00888	...	9.99193e-01	5.55112e-17	8.07103e-04	-1.94289e-16	0.99516	0.00484	1.61421e-02	7.26392e-03	0.94512	3.14770e-02
4	0.09203	0.43286	0.52711	0.50212	0.04914	0.05353	0.06727	0.04770	4.41696e-03	0.03004	...	9.98233e-01	1.76678e-03	0.00000e+00	0.00000e+00	0.98940	0.01060	1.76678e-01	8.83392e-03	0.77032	4.41696e-02
5	0.11691	0.64008	0.49449	0.48993	0.06097	0.06056	0.06942	0.08353	2.58095e-03	0.00469	...	1.00000e+00	-4.44089e-16	-2.81893e-18	-2.22045e-16	0.99437	0.00563	3.23792e-02	4.22337e-03	0.91741	4.59878e-02
6	0.03611	0.54779	0.49587	0.48299	0.02647	0.03776	0.04564	0.00343	-6.93889e-18	0.30735	...	9.32353e-01	6.76471e-02	-8.67362e-19	-4.85723e-17	0.07647	0.92353	1.47059e-03	7.35294e-03	0.77794	2.13235e-01
7	0.11832	0.62651	0.48841	0.49490	0.08230	0.08424	0.07652	0.15475	8.25083e-03	0.01430	...	9.96700e-01	3.30033e-03	-1.73472e-18	-1.45717e-16	0.99010	0.00990	1.76018e-02	1.10011e-03	0.92299	5.83058e-02
8	0.21769	0.56729	0.48480	0.48523	0.06343	0.07379	0.06494	0.00436	1.86916e-03	0.02617	...	9.60748e-01	9.34579e-03	2.99065e-02	2.77556e-17	0.54019	0.45981	1.86916e-03	3.73832e-03	0.00561	9.88785e-01
9	0.21501	0.39487	0.58130	0.49716	0.03949	0.05179	0.06317	0.00479	1.53846e-03	0.00718	...	9.50769e-01	4.92308e-02	-1.95156e-18	-1.52656e-16	0.99487	0.00513	1.02564e-02	6.07153e-18	0.05026	9.39487e-01
10	0.11946	0.56229	0.50303	0.48543	0.08458	0.08477	0.07114	0.01284	2.83126e-03	0.03171	...	9.73952e-01	2.60476e-02	-1.51788e-18	1.35900e-02	0.97508	0.01133	8.60702e-02	9.06002e-03	0.80974	9.51302e-02
11	0.21837	0.32000	0.55280	0.51927	0.04705	0.04564	0.06581	0.00283	-6.93889e-18	0.02545	...	8.46061e-01	1.53939e-01	-1.51788e-18	6.25455e-01	0.26909	0.10545	1.00606e-01	-8.67362e-19	0.01333	8.86061e-01
12	0.13164	0.55534	0.51973	0.49224	0.08655	0.08803	0.06964	0.01196	8.39695e-03	0.02061	...	9.89313e-01	1.06870e-02	-2.38524e-18	-2.08167e-16	0.99542	0.00458	8.16794e-02	7.63359e-03	0.85115	5.95420e-02

	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	...	deposit_type_No Deposit	deposit_type_Non Refund	deposit_type_Refundable	agent_1	agent_listed_other	agent_unknown	customer_type_Contract	customer_type_Group	customer_type_Transient	customer_type_Transient-Party
0	0.03572	0.55167	0.48726	0.48273	0.02679	0.03742	0.04537	0.00152	-6.93889e-18	0.30547	...	9.31611e-01	6.83891e-02	-8.67362e-19	-3.46945e-17	7.29483e-02	9.27052e-01	1.51976e-03	7.59878e-03	0.78267	0.20821
1	0.11744	0.60303	0.49879	0.48426	0.05885	0.05928	0.06791	0.03067	2.47934e-03	0.00606	...	1.00000e+00	-3.33067e-16	-2.81893e-18	-2.56739e-16	9.95041e-01	4.95868e-03	7.27273e-02	4.95868e-03	0.87218	0.05014
2	0.28486	0.56383	0.48881	0.49164	0.04014	0.05268	0.06718	0.00022	1.12757e-17	0.00598	...	6.66134e-16	1.00000e+00	-2.38524e-18	2.59309e-01	6.14362e-01	1.26330e-01	4.12234e-02	1.56125e-17	0.95811	0.00066
3	0.11149	0.64780	0.49134	0.49772	0.06193	0.06326	0.07970	0.09731	2.44798e-03	0.00306	...	9.98776e-01	1.22399e-03	-2.81893e-18	-2.42861e-16	9.96940e-01	3.05998e-03	2.63158e-02	8.56793e-03	0.91065	0.05447
4	0.25801	0.35543	0.60472	0.50362	0.04742	0.06045	0.06759	0.00859	9.20810e-04	0.00921	...	9.13444e-01	8.65562e-02	-2.16840e-19	1.38778e-17	9.94475e-01	5.52486e-03	7.73481e-02	1.84162e-03	0.12155	0.79926
5	0.13249	0.51493	0.53141	0.48946	0.05162	0.05584	0.06458	0.00687	3.08960e-03	0.01648	...	9.98970e-01	1.02987e-03	-1.73472e-18	-1.59595e-16	9.96910e-01	3.08960e-03	1.44181e-02	3.08960e-03	0.46859	0.51390
6	0.04617	0.51988	0.49986	0.50667	0.03946	0.04970	0.06680	0.04096	1.02410e-02	0.12169	...	9.73494e-01	2.65060e-02	-1.73472e-18	-1.24900e-16	-1.11022e-16	1.00000e+00	-4.16334e-17	6.02410e-03	0.90000	0.09398
7	0.09580	0.65840	0.49753	0.51671	0.06290	0.06711	0.07162	0.08015	1.90840e-02	0.01145	...	1.00000e+00	1.11022e-16	-1.30104e-18	-1.04083e-16	1.00000e+00	8.32667e-17	3.81679e-03	1.65394e-02	0.90712	0.07252
8	0.11430	0.55222	0.50920	0.49308	0.07816	0.07899	0.06912	0.01138	8.19113e-03	0.02526	...	9.90444e-01	9.55631e-03	-2.60209e-18	-2.28983e-16	9.94539e-01	5.46075e-03	4.64164e-02	7.50853e-03	0.88532	0.06075
9	0.21477	0.55340	0.49261	0.48291	0.06299	0.07379	0.06444	0.00583	1.94175e-03	0.02718	...	9.59223e-01	9.70874e-03	3.10680e-02	3.46945e-17	5.26214e-01	4.73786e-01	1.94175e-03	3.88350e-03	0.00388	0.99029
10	0.14875	0.58236	0.51623	0.48542	0.09575	0.09734	0.07294	0.01361	2.91545e-03	0.02187	...	9.66472e-01	3.35277e-02	-6.50521e-19	1.74927e-02	9.75219e-01	7.28863e-03	1.07872e-01	7.28863e-03	0.76676	0.11808
11	0.10522	0.72804	0.47129	0.47575	0.04865	0.05193	0.06927	0.00860	2.01450e-03	0.00886	...	9.99194e-01	5.55112e-17	8.05802e-04	-2.01228e-16	9.95971e-01	4.02901e-03	2.01450e-02	7.25222e-03	0.94118	0.03143
12	0.11969	0.63326	0.48536	0.49418	0.08294	0.08505	0.07665	0.15677	7.83875e-03	0.01456	...	9.96641e-01	3.35946e-03	-1.51788e-18	-1.31839e-16	9.89922e-01	1.00784e-02	1.79171e-02	1.11982e-03	0.92273	0.05823
13	0.19942	0.34620	0.54015	0.52120	0.04851	0.04627	0.06511	0.00305	-7.80626e-18	0.02749	...	9.14921e-01	8.50785e-02	-1.08420e-18	5.94241e-01	2.90576e-01	1.15183e-01	2.74869e-02	-3.46945e-18	0.01571	0.95681

	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	...	deposit_type_No Deposit	deposit_type_Non Refund	deposit_type_Refundable	agent_1	agent_listed_other	agent_unknown	customer_type_Contract	customer_type_Group	customer_type_Transient	customer_type_Transient-Party
0	0.11389	0.63968	0.49454	0.47741	0.06034	0.05982	0.06847	0.03965	2.68611e-03	0.00460	...	1.00000	-1.38778e-17	-2.16840e-18	-2.22045e-16	9.94628e-01	0.00537	3.60706e-02	2.30238e-03	0.91711	4.45127e-02
1	0.27779	0.58454	0.48284	0.49317	0.03982	0.05321	0.06687	0.00023	8.67362e-18	0.00619	...	0.00069	9.99313e-01	-2.81893e-18	2.21993e-01	6.35052e-01	0.14296	1.11022e-16	1.47451e-17	1.00000	-3.60822e-16
2	0.12795	0.55609	0.51888	0.49480	0.08448	0.08592	0.06974	0.01246	8.35322e-03	0.02228	...	0.98886	1.11376e-02	-2.38524e-18	-2.08167e-16	9.94431e-01	0.00557	5.25060e-02	7.95545e-03	0.88544	5.40971e-02
3	0.12506	0.57175	0.49842	0.48610	0.08482	0.08521	0.07115	0.01187	2.78087e-03	0.02892	...	0.97553	2.44716e-02	-1.73472e-18	1.33482e-02	9.69967e-01	0.01669	8.45384e-02	8.89878e-03	0.78865	1.17909e-01
4	0.06946	0.55179	0.49805	0.48214	0.03274	0.04485	0.04854	0.00317	1.19048e-03	0.24286	...	0.92976	5.59524e-02	1.42857e-02	-1.11022e-16	6.07143e-02	0.93929	1.19048e-03	5.95238e-03	0.61548	3.77381e-01
5	0.16660	0.41632	0.55459	0.49852	0.03816	0.05083	0.06206	0.00379	2.06612e-03	0.01653	...	0.95041	4.95868e-02	-2.16840e-18	-1.52656e-16	9.95868e-01	0.00413	1.44628e-02	2.06612e-03	0.18492	7.98554e-01
6	0.11457	0.69321	0.47428	0.53653	0.05899	0.06382	0.07546	0.31850	3.51288e-03	0.00703	...	0.99766	2.34192e-03	8.67362e-19	4.16334e-17	9.92974e-01	0.00703	4.91803e-02	9.36768e-03	0.87354	6.79157e-02
7	0.23236	0.17949	0.58396	0.49732	0.05043	0.04474	0.06825	0.00171	-6.07153e-18	0.03590	...	0.77436	2.25641e-01	-4.33681e-19	8.90598e-01	-1.11022e-16	0.10940	1.41880e-01	-6.07153e-18	0.02564	8.32479e-01
8	0.11108	0.56621	0.50737	0.49058	0.05892	0.05998	0.06946	0.02668	3.45850e-03	0.00692	...	1.00000	1.94289e-16	-1.95156e-18	-1.52656e-16	9.92095e-01	0.00791	9.28854e-02	5.92885e-03	0.87451	2.66798e-02
9	0.07641	0.59978	0.48540	0.51947	0.05928	0.06424	0.07154	0.09380	1.90583e-02	0.07623	...	0.99888	1.12108e-03	-1.51788e-18	-1.38778e-16	5.15695e-01	0.48430	-2.08167e-17	1.12108e-02	0.87892	1.09865e-01
10	0.10319	0.72340	0.47273	0.47801	0.04803	0.05109	0.06906	0.00567	1.70213e-03	0.00936	...	0.99915	1.11022e-16	8.51064e-04	-1.87350e-16	9.95745e-01	0.00426	1.95745e-02	7.65957e-03	0.94043	3.23404e-02
11	0.11217	0.67129	0.48541	0.48850	0.06372	0.06434	0.08205	0.01223	2.03915e-03	0.00326	...	0.99918	8.15661e-04	-2.38524e-18	-2.01228e-16	9.96737e-01	0.00326	1.30506e-02	8.97227e-03	0.92822	4.97553e-02
12	0.05979	0.58368	0.50194	0.49225	0.04305	0.05214	0.06659	0.01890	8.29876e-03	0.06362	...	0.99447	5.53250e-03	-1.08420e-18	-6.93889e-17	4.57815e-01	0.54219	4.14938e-03	1.10650e-02	0.90180	8.29876e-02
13	0.25141	0.53333	0.52587	0.51524	0.05925	0.06513	0.06607	0.00533	1.33333e-03	0.01200	...	0.98933	5.33333e-03	5.33333e-03	-8.32667e-17	9.18667e-01	0.08133	5.60000e-02	2.66667e-03	0.00533	9.36000e-01
14	0.12373	0.58701	0.50204	0.49142	0.08504	0.08625	0.07496	0.16982	8.82724e-03	0.01261	...	0.99622	3.78310e-03	-1.08420e-18	-9.02056e-17	9.87390e-01	0.01261	2.01765e-02	-2.60209e-18	0.89912	8.07062e-02

	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	...	deposit_type_No Deposit	deposit_type_Non Refund	deposit_type_Refundable	agent_1	agent_listed_other	agent_unknown	customer_type_Contract	customer_type_Group	customer_type_Transient	customer_type_Transient-Party
0	0.15059	0.59438	0.50903	0.49313	0.09717	0.09802	0.07317	0.01393	2.88184e-03	2.30548e-02	...	0.98271	1.72911e-02	-6.50521e-19	1.44092e-03	0.97839	0.02017	1.06628e-01	7.20461e-03	0.73487	1.51297e-01
1	0.12247	0.63874	0.49670	0.48864	0.06405	0.06233	0.07102	0.05220	2.74725e-03	4.16334e-17	...	1.00000	6.93889e-17	-1.08420e-18	-7.63278e-17	0.99588	0.00412	3.43407e-02	6.86813e-03	0.91896	3.98352e-02
2	0.21407	0.54391	0.51746	0.46752	0.04046	0.05482	0.06500	0.00189	7.08215e-04	2.54958e-02	...	0.21530	7.84703e-01	-1.08420e-18	-6.93889e-17	0.96742	0.03258	-3.46945e-17	-5.20417e-18	0.99858	1.41643e-03
3	0.21701	0.40145	0.56364	0.49771	0.04088	0.05244	0.06302	0.00451	1.44928e-03	6.76329e-03	...	0.95266	4.73430e-02	-2.16840e-18	-1.66533e-16	0.98841	0.01159	1.44928e-02	8.67362e-18	0.03865	9.46860e-01
4	0.11219	0.67196	0.48516	0.48894	0.06372	0.06430	0.08204	0.01222	2.03749e-03	3.25998e-03	...	0.99919	8.14996e-04	-2.16840e-18	-1.87350e-16	0.99674	0.00326	1.30399e-02	8.96496e-03	0.92747	5.05297e-02
5	0.08220	0.46374	0.51651	0.50108	0.04747	0.05157	0.06637	0.05153	3.81679e-03	2.29008e-02	...	0.99809	1.90840e-03	2.16840e-19	2.08167e-17	0.98855	0.01145	1.69847e-01	9.54198e-03	0.76908	5.15267e-02
6	0.12289	0.65190	0.48357	0.50422	0.08323	0.08509	0.07600	0.12321	8.22785e-03	1.39241e-02	...	0.99620	3.79747e-03	-1.51788e-18	-9.71445e-17	0.98608	0.01392	2.02532e-02	1.26582e-03	0.90000	7.84810e-02
7	0.05836	0.54290	0.51418	0.46756	0.03125	0.04179	0.04761	0.00357	6.70241e-04	2.68097e-01	...	0.95442	4.55764e-02	-1.08420e-18	-7.63278e-17	0.06971	0.93029	1.34048e-03	6.70241e-03	0.67158	3.20375e-01
8	0.08411	0.57306	0.47882	0.52854	0.06032	0.06810	0.07024	0.05707	1.85387e-02	7.96074e-02	...	0.98691	2.18103e-03	1.09051e-02	-1.45717e-16	0.40131	0.59869	3.27154e-03	1.19956e-02	0.78299	2.01745e-01
9	0.11135	0.73831	0.47266	0.47385	0.04921	0.05270	0.06980	0.00764	1.83318e-03	5.49954e-03	...	0.99908	1.80411e-16	9.16590e-04	-1.80411e-16	0.99633	0.00367	1.74152e-02	8.24931e-03	0.94042	3.39138e-02
10	0.30327	0.56206	0.47671	0.51322	0.04005	0.05111	0.06821	0.00034	-5.20417e-18	2.77556e-17	...	0.00101	9.98991e-01	-1.73472e-18	4.04642e-01	0.39758	0.19778	6.25631e-02	7.80626e-18	0.93744	-5.55112e-17
11	0.20038	0.37118	0.53652	0.51808	0.05000	0.04826	0.06558	0.00196	-6.93889e-18	3.05882e-02	...	0.91882	7.64706e-02	4.70588e-03	5.34118e-01	0.32941	0.13647	2.47059e-02	2.60209e-18	0.01176	9.63529e-01
12	0.10249	0.49746	0.50355	0.48227	0.07196	0.07223	0.06866	0.06658	1.01781e-02	4.58015e-02	...	0.99746	2.54453e-03	-1.51788e-18	-9.71445e-17	0.99109	0.00891	1.01781e-02	1.14504e-02	0.94148	3.68957e-02
13	0.11444	0.63877	0.49244	0.48854	0.05962	0.05997	0.06867	0.09780	2.75482e-03	5.50964e-03	...	1.00000	-1.52656e-16	-2.38524e-18	-2.35922e-16	0.99518	0.00482	3.78788e-02	2.75482e-03	0.90840	5.09642e-02
14	0.14320	0.57215	0.52300	0.48804	0.08543	0.08829	0.07039	0.00881	4.57317e-03	8.13008e-03	...	0.98374	1.42276e-02	2.03252e-03	-1.73472e-16	0.99390	0.00610	9.95935e-02	6.09756e-03	0.71138	1.82927e-01
15	0.06594	0.61671	0.50417	0.49307	0.04177	0.04984	0.06727	0.05740	8.92857e-03	5.48469e-02	...	0.99490	5.10204e-03	-1.30104e-18	-1.04083e-16	0.54464	0.45536	-3.46945e-17	8.92857e-03	0.91582	7.52551e-02

	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	...	deposit_type_No Deposit	deposit_type_Non Refund	deposit_type_Refundable	agent_1	agent_listed_other	agent_unknown	customer_type_Contract	customer_type_Group	customer_type_Transient	customer_type_Transient-Party
0	0.11241	0.67143	0.48542	0.48838	0.06383	0.06435	0.08212	1.22449e-02	2.04082e-03	3.26531e-03	...	9.99184e-01	8.16327e-04	-2.38524e-18	-2.01228e-16	9.98367e-01	1.63265e-03	1.30612e-02	9.79592e-03	9.27347e-01	0.04980
1	0.12212	0.60253	0.51258	0.49683	0.08181	0.08224	0.07245	8.68157e-02	7.28863e-03	9.71817e-03	...	9.88338e-01	1.16618e-02	-1.95156e-18	-1.66533e-16	9.94169e-01	5.83090e-03	5.83090e-02	3.88727e-03	8.41594e-01	0.09621
2	0.10356	0.72508	0.46936	0.47536	0.04822	0.05158	0.06915	6.64452e-03	1.66113e-03	9.96678e-03	...	9.99169e-01	8.32667e-17	8.30565e-04	-1.87350e-16	9.95847e-01	4.15282e-03	2.15947e-02	7.47508e-03	9.38538e-01	0.03239
3	0.18102	0.50463	0.53194	0.50278	0.04514	0.05405	0.06044	5.40123e-03	1.15741e-03	1.04167e-02	...	9.94213e-01	2.31481e-03	3.47222e-03	-1.17961e-16	9.93056e-01	6.94444e-03	1.27315e-02	3.47222e-03	6.66134e-16	0.98380
4	0.04135	0.53101	0.48909	0.49549	0.04066	0.05028	0.06709	4.30380e-02	1.07595e-02	1.26582e-01	...	9.94937e-01	5.06329e-03	-1.51788e-18	-1.04083e-16	0.00000e+00	1.00000e+00	-3.46945e-17	5.06329e-03	8.98734e-01	0.09620
5	0.10962	0.55724	0.51032	0.49736	0.05821	0.05881	0.06965	5.04225e-02	3.67948e-03	1.22649e-02	...	1.00000e+00	5.55112e-17	-2.38524e-18	-2.01228e-16	9.95094e-01	4.90597e-03	8.83074e-02	7.35895e-03	8.86345e-01	0.01799
6	0.37484	0.47044	0.55824	0.49437	0.03948	0.04949	0.06907	2.77556e-17	-6.07153e-18	4.16334e-17	...	1.68919e-03	9.98311e-01	-4.33681e-19	7.11149e-01	0.00000e+00	2.88851e-01	1.04730e-01	-5.20417e-18	8.61486e-01	0.03378
7	0.13072	0.64387	0.47879	0.49964	0.08603	0.08912	0.07519	1.42096e-02	4.44050e-03	1.42096e-02	...	9.94671e-01	5.32860e-03	-2.16840e-19	2.08167e-17	9.82238e-01	1.77620e-02	2.84192e-02	1.77620e-03	8.89876e-01	0.07993
8	0.23330	0.61099	0.45253	0.48407	0.03898	0.05507	0.06581	3.52361e-04	-5.20417e-18	7.39958e-03	...	2.10942e-15	1.00000e+00	-1.73472e-18	-1.45717e-16	9.75687e-01	2.43129e-02	-1.38778e-17	5.20417e-18	9.98943e-01	0.00106
9	0.10202	0.50552	0.50297	0.48630	0.07224	0.07316	0.06885	7.48466e-02	1.04294e-02	4.53988e-02	...	9.97546e-01	2.45399e-03	-1.30104e-18	-1.17961e-16	9.92638e-01	7.36196e-03	9.81595e-03	9.81595e-03	9.48466e-01	0.03190
10	0.19729	0.21084	0.56271	0.49772	0.05422	0.04558	0.06723	2.00803e-03	-5.20417e-18	4.21687e-02	...	9.09639e-01	9.03614e-02	2.16840e-19	8.71486e-01	1.11022e-16	1.28514e-01	4.21687e-02	-4.33681e-18	2.00803e-02	0.93775
11	0.10830	0.54100	0.50727	0.50969	0.03699	0.05495	0.05348	3.56506e-03	1.78253e-03	1.35472e-01	...	8.98396e-01	7.84314e-02	2.31729e-02	2.08167e-17	6.95187e-02	9.30481e-01	1.78253e-03	5.34759e-03	4.90196e-01	0.50267
12	0.03630	0.54408	0.50354	0.48010	0.03212	0.03520	0.04618	1.67926e-03	-2.60209e-18	3.29975e-01	...	9.47103e-01	5.28967e-02	8.67362e-19	4.85723e-17	7.30479e-02	9.26952e-01	-1.38778e-17	7.55668e-03	6.87657e-01	0.30479
13	0.25697	0.39519	0.56450	0.50342	0.05008	0.06262	0.06723	6.21118e-03	1.55280e-03	7.76398e-03	...	9.25466e-01	7.45342e-02	-6.50521e-19	-2.77556e-17	9.89130e-01	1.08696e-02	6.52174e-02	1.55280e-03	8.69565e-02	0.84627
14	0.09666	0.65816	0.49715	0.51488	0.06322	0.06738	0.07171	8.03571e-02	1.91327e-02	1.14796e-02	...	1.00000e+00	1.24900e-16	-1.30104e-18	-9.02056e-17	1.00000e+00	1.11022e-16	3.82653e-03	1.65816e-02	9.05612e-01	0.07398
15	0.15083	0.59294	0.51042	0.49207	0.09753	0.09820	0.07333	1.39289e-02	2.88184e-03	2.30548e-02	...	9.82709e-01	1.72911e-02	-1.08420e-18	1.44092e-03	9.81268e-01	1.72911e-02	1.06628e-01	7.20461e-03	7.37752e-01	0.14841
16	0.11495	0.63889	0.49554	0.48828	0.05954	0.05977	0.06878	9.75610e-02	2.71003e-03	5.42005e-03	...	1.00000e+00	-1.38778e-16	-2.60209e-18	-2.35922e-16	9.92547e-01	7.45257e-03	3.65854e-02	2.03252e-03	9.11924e-01	0.04946

	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	...	deposit_type_No Deposit	deposit_type_Non Refund	deposit_type_Refundable	agent_1	agent_listed_other	agent_unknown	customer_type_Contract	customer_type_Group	customer_type_Transient	customer_type_Transient-Party
0	0.15755	0.61502	0.52372	0.46454	0.10803	0.10359	0.07443	1.04366e-01	4.79233e-03	1.91693e-02	...	0.97444	2.55591e-02	8.67362e-19	3.46945e-17	0.99361	0.00639	8.30671e-02	9.58466e-03	7.47604e-01	1.59744e-01
1	0.12374	0.53590	0.51438	0.48424	0.08417	0.08213	0.06974	1.19658e-02	5.12821e-03	3.38462e-02	...	1.00000	2.08167e-16	-1.73472e-18	-1.59595e-16	0.99795	0.00205	1.23077e-01	9.23077e-03	8.55385e-01	1.23077e-02
2	0.07572	0.60267	0.48769	0.52146	0.05873	0.06383	0.07154	9.55143e-02	1.97216e-02	7.42459e-02	...	1.00000	1.52656e-16	-1.30104e-18	-1.24900e-16	0.52668	0.47332	-2.77556e-17	1.16009e-02	9.04872e-01	8.35267e-02
3	0.19829	0.35325	0.53866	0.52130	0.04821	0.04601	0.06503	1.73160e-03	-6.93889e-18	2.72727e-02	...	0.91558	8.44156e-02	-1.30104e-18	5.89610e-01	0.30260	0.10779	2.72727e-02	-2.60209e-18	1.29870e-02	9.59740e-01
4	0.10453	0.72353	0.47519	0.47686	0.04827	0.05141	0.06909	6.72269e-03	1.68067e-03	9.24370e-03	...	0.99916	8.32667e-17	8.40336e-04	-1.94289e-16	0.99580	0.00420	1.84874e-02	7.56303e-03	9.41176e-01	3.27731e-02
5	0.11242	0.67075	0.48647	0.48813	0.06367	0.06440	0.08206	1.22549e-02	2.04248e-03	3.26797e-03	...	0.99918	8.16993e-04	-2.16840e-18	-2.01228e-16	0.99673	0.00327	1.30719e-02	8.98693e-03	9.27288e-01	5.06536e-02
6	0.24658	0.55435	0.52094	0.45556	0.03770	0.05611	0.06421	2.08167e-17	-6.07153e-18	1.26812e-02	...	0.00000	1.00000e+00	-2.16840e-19	1.38778e-17	0.95833	0.04167	-2.08167e-17	-5.20417e-18	9.98188e-01	1.81159e-03
7	0.11964	0.58736	0.49190	0.49703	0.07977	0.08138	0.07490	1.35894e-01	4.95663e-03	1.61090e-02	...	0.99504	4.95663e-03	-1.30104e-18	1.23916e-03	0.99380	0.00496	6.31970e-02	4.95663e-03	8.27757e-01	1.04089e-01
8	0.10784	0.57585	0.51016	0.49966	0.06689	0.07193	0.06888	1.10079e-02	1.03199e-02	2.27038e-02	...	0.99794	2.06398e-03	-1.51788e-18	-1.52656e-16	0.99174	0.00826	0.00000e+00	5.15996e-03	9.16409e-01	7.84314e-02
9	0.06011	0.58414	0.50342	0.49191	0.04310	0.05207	0.06662	1.88506e-02	8.27586e-03	6.34483e-02	...	0.99448	5.51724e-03	-1.08420e-18	-6.24500e-17	0.45793	0.54207	4.13793e-03	1.10345e-02	9.04828e-01	8.00000e-02
10	0.11837	0.61310	0.49954	0.54355	0.06101	0.06391	0.07187	3.95833e-01	4.46429e-03	5.95238e-03	...	0.99702	2.97619e-03	8.67362e-19	4.16334e-17	0.99405	0.00595	6.25000e-02	8.92857e-03	8.48214e-01	8.03571e-02
11	0.22079	0.38963	0.58458	0.50710	0.03874	0.05028	0.06264	2.60126e-03	1.11483e-03	4.45931e-03	...	0.94760	5.23969e-02	-1.73472e-18	-1.24900e-16	0.99443	0.00557	1.11483e-03	3.46945e-18	8.88178e-16	9.98885e-01
12	0.21198	0.54910	0.49491	0.48029	0.06212	0.07320	0.06443	6.01202e-03	3.00601e-03	2.60521e-02	...	0.95792	1.00200e-02	3.20641e-02	3.46945e-17	0.50701	0.49299	2.00401e-03	2.00401e-03	1.00200e-02	9.85972e-01
13	0.11067	0.76797	0.30894	0.48571	0.05784	0.05732	0.06822	3.93375e-02	3.10559e-03	5.32387e-03	...	1.00000	1.38778e-16	-2.16840e-18	-1.80411e-16	0.99645	0.00355	4.16334e-17	1.77462e-03	9.39663e-01	5.85626e-02
14	0.12075	0.37776	0.74557	0.47761	0.05952	0.06153	0.06868	3.41482e-02	2.32829e-03	5.82072e-03	...	1.00000	1.66533e-16	-1.51788e-18	-1.17961e-16	0.98952	0.01048	1.64144e-01	8.14901e-03	7.84633e-01	4.30733e-02
15	0.30347	0.54666	0.48400	0.51081	0.04009	0.05169	0.06807	9.02056e-17	-3.46945e-18	2.08167e-17	...	0.00098	9.99018e-01	-1.95156e-18	3.93910e-01	0.39784	0.20825	6.09037e-02	8.67362e-18	9.39096e-01	-5.55112e-17
16	0.12982	0.64646	0.47991	0.49596	0.08722	0.09063	0.07559	1.43035e-02	4.66418e-03	1.49254e-02	...	0.99440	5.59701e-03	0.00000e+00	2.77556e-17	0.98694	0.01306	2.98507e-02	1.86567e-03	9.06716e-01	6.15672e-02
17	0.02842	0.55805	0.48905	0.48179	0.02757	0.03611	0.04541	2.57998e-03	-6.93889e-18	3.14241e-01	...	0.96594	3.40557e-02	-6.50521e-19	-2.77556e-17	0.08359	0.91641	1.54799e-03	7.73994e-03	7.67802e-01	2.22910e-01

	lead_time	arrival_date_year	arrival_date_week_number	arrival_date_day_of_month	stays_in_weekend_nights	stays_in_week_nights	adults	children	babies	is_repeated_guest	...	deposit_type_No Deposit	deposit_type_Non Refund	deposit_type_Refundable	agent_1	agent_listed_other	agent_unknown	customer_type_Contract	customer_type_Group	customer_type_Transient	customer_type_Transient-Party
0	0.30861	0.52746	0.48632	0.51158	0.03954	0.05047	0.06843	0.00032	-1.73472e-18	6.93889e-18	...	0.00095	9.99053e-01	-1.95156e-18	4.41288e-01	3.73106e-01	1.85606e-01	5.87121e-02	9.54098e-18	0.87973	0.06155
1	0.22142	0.38933	0.58364	0.50176	0.03933	0.05171	0.06334	0.00337	1.68539e-03	8.98876e-03	...	0.94719	5.28090e-02	-1.95156e-18	-1.31839e-16	9.95506e-01	4.49438e-03	7.86517e-03	1.12360e-03	0.04045	0.95056
2	0.04860	0.53341	0.49336	0.52116	0.04329	0.05217	0.06789	0.06088	1.22494e-02	1.49220e-01	...	1.00000	-6.93889e-17	4.33681e-19	4.85723e-17	1.11022e-16	1.00000e+00	-1.38778e-17	2.22717e-03	0.88641	0.11136
3	0.10523	0.72888	0.47121	0.47664	0.04862	0.05187	0.06925	0.00885	2.01126e-03	8.84956e-03	...	0.99920	4.16334e-17	8.04505e-04	-2.01228e-16	9.95977e-01	4.02253e-03	1.85036e-02	7.24055e-03	0.94288	0.03138
4	0.14162	0.57658	0.51698	0.49404	0.09769	0.09779	0.07352	0.01401	3.00300e-03	2.25225e-02	...	0.98198	1.80180e-02	-8.67362e-19	1.50150e-03	9.90991e-01	7.50751e-03	1.11111e-01	7.50751e-03	0.77628	0.10511
5	0.13193	0.58114	0.51811	0.48968	0.08550	0.08657	0.07086	0.00952	5.14286e-03	8.00000e-03	...	0.98629	1.37143e-02	-1.95156e-18	-1.31839e-16	9.93143e-01	6.85714e-03	1.10857e-01	6.85714e-03	0.78857	0.09371
6	0.10221	0.50825	0.50069	0.49033	0.07325	0.07442	0.06766	0.03668	1.16919e-02	4.67675e-02	...	0.99725	2.75103e-03	-1.08420e-18	-6.93889e-17	9.90371e-01	9.62861e-03	1.65062e-02	1.10041e-02	0.92985	0.04264
7	0.22067	0.54579	0.51917	0.46972	0.03997	0.05535	0.06494	0.00148	-6.93889e-18	1.92024e-02	...	0.18464	8.15362e-01	-1.08420e-18	-5.55112e-17	9.66027e-01	3.39734e-02	-3.46945e-17	-5.20417e-18	0.99852	0.00148
8	0.12914	0.64428	0.47690	0.49800	0.08621	0.09020	0.07549	0.01452	4.53721e-03	1.45191e-02	...	0.99456	5.44465e-03	0.00000e+00	2.08167e-17	9.87296e-01	1.27042e-02	2.90381e-02	1.81488e-03	0.90926	0.05989
9	0.18352	0.38495	0.53192	0.52793	0.05056	0.04682	0.06464	0.00193	-6.93889e-18	3.03907e-02	...	1.00000	6.93889e-17	-8.67362e-19	5.62952e-01	3.40087e-01	9.69609e-02	3.03907e-02	1.44718e-03	0.01592	0.95224
10	0.08070	0.45437	0.53571	0.48810	0.03472	0.04812	0.05306	0.00794	8.67362e-19	5.55556e-02	...	0.92857	7.14286e-02	4.33681e-19	4.16334e-17	1.66667e-01	8.33333e-01	3.96825e-03	7.93651e-03	0.28571	0.70238
11	0.22331	0.56713	0.48520	0.48444	0.06300	0.07350	0.06517	0.00401	3.00601e-03	2.40481e-02	...	0.95792	1.00200e-02	3.20641e-02	3.46945e-17	5.39078e-01	4.60922e-01	2.00401e-03	2.00401e-03	0.00601	0.98998
12	0.11542	0.60355	0.50026	0.48667	0.05933	0.05945	0.06821	0.03283	2.79188e-03	6.09137e-03	...	1.00000	-4.02456e-16	-2.81893e-18	-2.70617e-16	9.95431e-01	4.56853e-03	6.64975e-02	4.06091e-03	0.88528	0.04416
13	0.09300	0.48341	0.54785	0.45340	0.05895	0.06049	0.07074	0.03476	3.55450e-03	1.18483e-02	...	1.00000	-8.32667e-17	8.67362e-19	3.46945e-17	9.92891e-01	7.10900e-03	4.26540e-02	2.36967e-02	0.78436	0.14929
14	0.09663	0.64681	0.49484	0.51553	0.06955	0.07261	0.07439	0.12057	2.44681e-02	1.70213e-02	...	1.00000	-8.32667e-17	6.50521e-19	3.46945e-17	1.00000e+00	-1.66533e-16	-2.08167e-17	1.91489e-02	0.91702	0.06383
15	0.06312	0.59398	0.49569	0.49719	0.04492	0.05297	0.06695	0.01805	9.02256e-03	4.66165e-02	...	0.99398	6.01504e-03	-8.67362e-19	-4.85723e-17	4.69173e-01	5.30827e-01	4.51128e-03	1.20301e-02	0.90075	0.08271
16	0.11315	0.69172	0.47720	0.49012	0.06361	0.06375	0.08246	0.01182	2.11149e-03	3.37838e-03	...	0.99916	8.44595e-04	-2.38524e-18	-1.94289e-16	9.97466e-01	2.53378e-03	1.35135e-02	1.68919e-03	0.93412	0.05068
17	0.01334	0.59091	0.47250	0.46800	0.02370	0.03243	0.04221	0.00148	-4.33681e-18	4.27938e-01	...	0.95787	4.21286e-02	8.67362e-19	4.16334e-17	4.43459e-02	9.55654e-01	-2.08167e-17	6.65188e-03	0.98448	0.00887
18	0.11776	0.62875	0.49593	0.52046	0.06779	0.06966	0.07466	0.42916	7.93651e-03	8.81834e-03	...	0.99824	1.76367e-03	0.00000e+00	1.38778e-17	9.98236e-01	1.76367e-03	3.52734e-02	5.29101e-03	0.88360	0.07584

DePaul University Jarvis College of Computing and Digital Media¶

DSC 478 Final Project¶

Outline¶

1 - Data Collection:¶

1.1 - Dataset: Hotel booking demand¶

1.2 - Importing the Data:¶

2 - Data Preprocessing:¶

2.1 - Summary Statistics:¶

2.2 - Data Cleaning:¶

2.3 - Data Visualization and Exploration:¶

2.4 - Feature Engineering:¶

I can now safely drop the columns:¶

2.5 Data Visualization for Target Variable¶

Preparing the data and labels for classification¶

3 - Supervised Knowledge Discovery¶

3.1 KNN¶

First split, then train the minmax scalar on train set only; then use it to transform test:¶

3.2 Decision Trees¶

3.3 SVM¶

3.4 Naive Bayes¶

3.5 linear discriminant analysis¶

3.6 Random Forest¶

4 - Unsupervised Knowledge Discovery¶

4.1 Clustering¶

4.2 Qualitative Analysis of Clusters¶

Detecting patterns among cluster centroids:¶

5 - Comparison and Conclusion¶

KNN : 0.80(+/- 0.01)¶

Decision Trees : 0.82 (+/- 0.02)¶

Random Forest : : 0.85 (+/- 0.02)¶

SVM: 0.79 (+/- 0.01)¶

LDA: 0.78 (+/- 0.02)¶

Naïve Bayes: 0.53 (+/- 0.03)¶

1.1 - Dataset: Hotel booking demand ¶